IODP Proceedings    Volume contents     Search

doi:10.2204/iodp.proc.324.106.2010

Igneous petrology

Stratigraphic unit description and volcanology

In this site report, stratigraphic units are given Roman numbering (Units I–V), whereas the small-scale lithologic units identified during core description are given Arabic numbering (Units 1–32). We identified these lithologic units on the basis of sedimentological changes or, in the case of volcanological units, criteria including the presence of chilled margins or contacts, identifiable flow tops, vesicle distribution, and the occurrence of intercalated volcaniclastic or sedimentary horizons. When considered together, these criteria define the various sizes of individual volcanic inflation units. Accordingly, in the volcanic context, the term "stratigraphic unit" is used to combine consecutive inflation units of similar character into lava packages or eruptive successions (see "Igneous petrology" in the "Methods" chapter).

Hole U1349A was drilled near the top of Ori Massif, a large volcanic edifice the size of the Big Island of Hawaii (Fig. F1). The uppermost succession of recovered sediment commences with an Lower Cretaceous sequence consisting of pelagic red chert and nannofossil ooze (stratigraphic Unit I) followed by low recovery (~2%) of clay-, silt-, and sandstone and fragments of lithified calcareous sandstone (stratigraphic Unit II), which also includes a component of granular volcaniclastic material (Fig. F5). Beneath these upper units, sediment recovery improves greatly (~62% beginning with Core 324-U1349A-5R) and reveals a complex succession (stratigraphic Unit III) that, in its upper part, begins with ~8 m of gray-green volcaniclastic sediment containing normally graded beds (each beginning with gravel-sized particles), which may have been deposited as turbidites. These pass downward into Subunit IIIa, consisting of layered, coarse volcanic sandstone containing abundant pebble-sized clasts of basalt and hyaloclastite; this subunit has reddish (ferric) iron staining in its lower part. Subunit IIIb consists of ~1 m of reddish orange oligomict volcanogenic conglomerate. A marked reddened horizon in the upper part of this subunit is interpreted as a possible paleosol and may indicate oxidative alteration in a subaerial eruptive environment. Beneath this horizon, the conglomerate is weathered brown throughout, becomes coarser, contains subrounded cobble-sized volcanic clasts (including magnetite- and plagioclase-rich examples), and lies unconformably upon a highly vesicular weathered basalt flow (i.e., the top of stratigraphic Subunit IVa). For information on the sediments in Units I–III (Fig. F5) see "Sedimentology."

Volcanic basement was reached at 165.1 mbsf (Sample 324-U1349A-7R-1, 96 cm) (Fig. F13). Drilling continued into a ~54 m succession of highly altered (~20%–60%) vesicular lava flows intercalated with thin beds of carbonate-rich sand- and mudstones (Unit IV) and underlain by ~29 m of flow breccias with abundant altered olivine and (relatively) higher MgO contents (Unit V). The actual contact between the reddish oligomict conglomerate and the topmost basalt flow was not recovered, but the majority of flow units in the underlying vesicular lava succession (Unit IV) exhibit a brownish orange ferric alteration, as well as a number of deeply reddened (possibly weathered) flow tops. In addition to weathering horizons, several flows are intercalated with carbonate-rich sedimentary horizons.

The following description first outlines Unit IV in terms of an upper lava succession, consisting of thin, highly vesicular lava flow units (Subunit IVa), and a lower lava succession, characterized by thicker lava inflation units (Subunit IVc) affected by pervasive oxidative weathering; Subunits IVa and IVc are distinguished by an intervening layer of oolitic limestone (Subunit IVb). Second, the description outlines Unit V, the underlying volcanic flow breccias and massive olivine-phyric picritic lava pods, which have been affected by a different gray-green clay-rich alteration.

Vesicular lava succession (stratigraphic Unit IV, lithologic Units 4–31)

A total of ~54 m of highly altered, ~0.5–6 m thick vesicular lava units were recovered in this succession (Figs. F13, F14). The succession may be divided into an upper (~5 m; Subunit IVa) and lower (~49 m; Subunit IVc) lava section, respectively lying above and below a ~6 m oolitic limestone intercalation (as determined from downhole logging data; Subunit IVb). Pahoehoe-like flow structures are preserved in the upper parts of the thinner flow units that occur in both Subunits IVa and IVc. These pahoehoe lava flow tops are characterized by a particularly high degree of vesicularity (>40%–75%) and display a range of structures such as lava mixing and brecciated vesicular upper crusts (Figs. F15, F16). These inflation units, with rubbly brecciated upper crusts and massive amygdaloidal interiors, are more clearly evident in cores with higher recovery, such as Core 324-U1349A-13R (Fig. F16). The vesicular texture of the upper lava sections is "spongelike," consisting of mainly horizontal and intensely vesiculated zones several tens of centimeters thick. In addition, these sections contain rare zones of stretched/distorted vesicles that help to reveal folds, brecciated zones, and partially digested pahoehoe crusts in the lava flow tops.

In addition to the ~6 m thick oolitic limestone intercalation, Subunit IVb, many other smaller intercalations of limestone containing sand-size peloids and oolites occur throughout Subunit IVa and the topmost part of Subunit IVc. These limestone deposits often fill voids and vesicles in the uppermost few centimeters of the lava flows, these being generally highly vesicular and often reddened. In these intervals the usual calcitic or micritic vesicle infilling is absent, and infilling is dominated by clays instead.

X-ray diffraction (XRD) data reveal mostly Fe oxyhydroxides (low-crystallinity hematite and goethite), together with montmorillonite, halloysite, and traces of some primary igneous phases such as feldspar (interval 324-U1349A-10R-1, 1–35 cm). In these reddened lava-top horizons, this combination of secondary minerals is consistent with low-temperature subaerial alteration.

Thin sections reveal that the lavas of Unit IV contain 3%–5% olivine phenocrysts set in a matrix that contains many "ophimottle" aggregates (see "Petrography and igneous petrology") of subophitic clinopyroxene and plagioclase. In addition, Sections 324-U1349A-9R-1 and 9R-2 (Subunit IVc) contain discrete vesicular horizons apparently disrupted by late-stage magma mingling and postemplacement injection and movement within the lava inflation units themselves. These zones may represent disrupted sections of partially or fully solidified crust between which a later, degassed magma appears to have been injected. In particular, the highly vesicular upper part of an ~8 m thick inflation unit in Core 324-U1349A-11R preserves a range of magma mixing features, showing intrusion and disruption of the surrounding, more vesiculated horizons (Fig. F15). This heterogeneous flow exhibits several mixing contacts consisting of sparsely vesicular, microcrystalline, and reddish orange basaltic lava intruding into a more highly vesicular and sparsely olivine-phyric basalt that evidently is of another lava type. However, inductively coupled plasma–atomic emission spectroscopy (ICP-AES) geochemical data indicate that these two lava types are essentially similar, suggesting that the observed differences in lava type are most probably caused by differences in vesicularity and degree of alteration rather than petrogenesis.

Thicker lava inflation units occur toward the bottom of the lower lava section (Subunit IVc). For example, improved recovery (~60%–80%) in Cores 324-U1349A-10R through 12R shows that this subunit contains three lava inflation units that are between 4 and 6 m thick, whereas the lowermost succession within Core 13R consists of units that are between 1 and 3 m thick (Fig. F16). These thicker inflation units also have highly vesicular tops but are sparsely vesicular in their interiors and range in coloration from brown to gray-green depending upon the degree and style of alteration. Alteration is particularly complex in these lower massive units. For instance, lithologic Unit 22 (~6 m thick inflation unit) between intervals 324-U1349A-11R-2, 87 cm, and 11R-6, 129 cm, contains not only primary magma mixing features but also oxidative reddening horizons, a pervasive brown alteration color, abundant veining, and zeolite infilling along fractures and in vesicles (see "Alteration and metamorphic petrology").

The contact between the lowermost flow and the underlying flow breccia sequence (Unit V) is not preserved, but the alteration described above continues farther down, affecting the topmost part of the underlying unit by imparting a similar brownish color (Sections 324-U1349A-14R-1 to 14R-3).

Flow breccia sequence (stratigraphic Unit V, lithologic Unit 32)

Unit V consists of the interval from Section 324-U1349A-14R-1 to the bottom of hole. It is a ~25 m thick, poorly sorted alternating assemblage of roughly equal proportions of small, irregular lava inflation units (pods) and fragmental basalt breccia in which no internal bedding or grading can be discerned (Figs. F17, F18). In the core the discernible lava pods are vesicular and between 0.5 and 2 m thick (average ~1 m), which is probably an accurate representation of their true dimension because core recovery was very good (~75%–90%). The cores of these lava pods usually have spherical vesicles that grade into deformed, elongate vesicles along the outer margins. Their margins typically consist of fragmented basaltic lava crust that exhibits jigsaw-fit textures, thus indicating that relatively little transport occurred after their emplacement. The margins of these pods appear to have chilled against the matrix of basaltic breccia onto, or into, which they were emplaced. Glass shards are generally not easily discernible in the breccia because of alteration but are identifiable in the matrix of the fragmental horizons (see "Petrography and igneous petrology"). Fine granule- to sand-sized fragments, commonly occurring in pockets, are densely concentrated along the margins of the more disintegrated inflation units.

The fragmental areas of the flow breccias are typically between 0.1 and 1.3 m thick (average = 0.8 m). The constituent basalt fragments are mostly angular and several centimeters in size (0.1–10 cm). Several horizons consist of fractured, cobble-size basaltic fragments. All the brecciated horizons are monomict (i.e., they contain vesicular hyalobasalt clasts or glass shards), poorly sorted, and may be either clast or matrix supported depending upon the relative proportions of fine to coarse constituents. The clasts display a wide range of morphologies that range from blocky jigsaw-fit basalt clasts to irregular clasts with wavy and subplanar margins. Single basalt fragments exhibit mostly fragmented morphologies but preserve textural features that indicate they formed in situ (because adjacent fragments have contiguous shapes). Petrographic textures also reveal complex wavy and jigsaw-fit morphologies (Figs. F17, F18) at the micro scale, and many of the clasts with irregular morphologies (i.e., chipped hyaloclastite fragments) have thin dark green rims that may interpreted as chilled margins, typical of submarine eruptions. Therefore, Unit V is classified as a hyaloclastite pillow breccia.

Preliminary assessment

Varying eruption environments (stratigraphic Units IV and V)

The lava succession of Unit IV contains a variety of nonvolcanic features. These include reddening of many of the flow tops and evidence of shallow-water sedimentary intercalations. Oxidative alteration of lava tops is likely indicative of subaerial alteration and, coupled with the shallow-marine deposits preserved at the contacts of the uppermost lava units, suggests a littoral and/or periodically "emergent" environment during the development of the lava succession. Because the conglomerate of Subunit IIIb lies immediately on top of the lava succession and contains a putative paleosol within it, these features can be interpreted to represent a period of exposure and erosion following the emplacement of the lava succession. The variety of clast types within the conglomerate includes compositions that cannot be readily matched to the succession immediately below and so might indicate emergence and erosion of a volcanically complex terrain.

By contrast, Unit V is interpreted as an autoclastic volcanic breccia that involved fragmentation of semisolid or solid lava (Fisher, 1960). Flow breccia may be formed by nonexplosive fragmentation of flowing or recently solidified basaltic lava and preserves features that indicate disintegration and/or autobrecciation of coherent lava flows. This flow breccia is a distinct, heterogeneous unit with complex internal features that formed as part of a series of individual flows or injections into the unconsolidated hyaloclastite that underwent fragmentation. Importantly, the jigsaw-puzzle texture in many of the flow breccia fragments indicates there was minimal transportation after emplacement of the lava unit. The fracturing present within individual lava flows must, therefore, be caused by limited mobilization or mechanical friction within the unconsolidated materials, or else by contraction of lava pods during cooling. The highly fluidal shaped morphologies of some smaller clasts in the brecciated horizons suggest fragmentation took place while some parts of the flow were still partially molten. Such fragmentation appears to have taken place especially in the highly vesicular areas of the flow. Emplacement of these flow breccias could have resulted from autobrecciation processes in which effusive subaerial eruptions enter into submarine environments or, alternatively, gaseous explosions during submarine lava emplacement as part of a lava delta sequence, or possibly mass mobilization related to mass movement or seismic activity. However, the latter is least likely since many fragments indicate that postemplacement movement was minimal.

Accepting the above interpretations, the volcanic succession of Hole U1349A reveals a change from submarine eruption, producing a thick stack of autobrecciated lavas and hyaloclastite breccias, passing upward into a phase where thick inflation units developed (up to 6 m thick). These thick units have weathered surfaces, indicating that by this stage the succession had perhaps built itself above sea level. However, periodic marine incursion permitted deposition of carbonate sediments, which infilled voids, pockets, and vesicles within the exposed lava surfaces. The uppermost part of the lava succession therefore appears to be an interplay between subaerial exposure and very shallow marine conditions over a low-lying volcanic landscape. Further erosion then led to the development of the conglomerate, before reestablishment of marine conditions resulted in the deposition of the volcaniclastic sedimentary succession of Unit III.

Determination of eruption depth for the highly vesicular lavas (stratigraphic Unit IV)

One of the major questions relating to the volcanism at Shatsky Rise concerns the water depth in which the lavas were erupted. The high 40–75 vol% vesicularity or spongelike basalt lavas of Unit IV provide an opportunity to make an initial estimate of eruption depth using assumptions based on equilibrium degassing and the likely rapid quenching of outpouring magma. Vesicle content of submarine glasses recovered from known eruption depths on the eastern rift zone of Kilauea Volcano (Hawaii) has been shown to agree with vesicularities calculated from degassing models in which the pre-eruptive H2O and CO2 contents have been constrained (Wallace, 1998). This correlation also may be applied to the vesicularity measurements of the Unit IV lavas (Fig. F14) and thus used to estimate the eruption depths. Precise estimation would require determination of H2O and CO2 concentrations for the lavas, but a preliminary estimate is possible by using the degassing curves for different proportions of H2O and CO2 contents of basaltic melt provided by Wallace (1998) (Fig. F19). Thus, assuming that H2O concentrations in Shatsky Rise magma are typical for mid-ocean-ridge basalts (MORBs) elsewhere (i.e., 0.05–0.4 wt% H2O; Danyushevsky, 2001), the observed 40–75 vol% vesicularity suggests that the lavas could have formed at pressures below 15–20 bar, corresponding to water depths of less than 150–200 mbsl (Fig. F19). It should be emphasized, however, that estimated eruption pressures are highly dependent on the assumed initial melt H2O concentrations; if the magma initially were more hydrous, deeper eruptions would be predicted.

Petrography and igneous petrology

In Hole U1349A, true extrusive basement (Units IV and V) consists of ~84 m of mainly pillow lavas, massive flows, and flow top breccias. An overview of these volcanic basement units and the abundances of olivine and clinopyroxene phenocrysts in these rocks are shown in Figure F20 and Table T5. Polished thin sections were prepared of rims and interiors of 55 of these rocks, including all intervals where ICP-AES analyses were made. A summary table of thin section descriptions of the volcaniclastic rocks is given in 324TS.XLS in LOGS in "Supplementary material," and the contacts between units are described in 324UNIT.XLS in LOGS in "Supplementary material." The principal igneous lithologies identified in thin section are

  • Subunit IIIb: varieties of lithic, subaerial, and glassy submarine volcanic rocks found in a polymict conglomerate derived partly from detrital and partly from nearby active volcanic sources;

  • Subunit IVa: highly vesicular olivine-spinel basalt, extensively oxidized and altered, with textural aspects suggesting both subaerial and submarine extrusion;

  • Subunit IVc: generally fine- to medium-grained olivine-spinel phyric basalts, some of which contain unusually coarse patches of plagioclase-clinopyroxene intergrowths exhibiting ophitic texture and all of which are variably, but extensively, oxidized and altered; and

  • Unit V: dark khaki-green olivine-spinel phyric basalt and/or picrite that is extensively altered.

Despite the alteration, primary petrographic features of the basalts can be recognized. The following describes the rocks as they were recovered sequentially downhole and emphasizes their primary igneous features based on thin section observation. Identification of the least altered igneous minerals and the influence of alteration on igneous textures will also be covered. The most outstanding petrographic features of the rocks are

  • Occurrence of ferroandesite clasts of (likely) tholeiitic or (less likely) alkalic parentage in the oligomict volcanogenic conglomerate of Subunit IIIb;

  • Extraordinary vesicularity of the pillow lava crusts, up to ~75%, in rocks of Subunit IVa at the top of the basement, which together with interbedded sedimentary rock and a capping paleosol are likely to be indicative of eruption in very shallow water and, in part, subaerially;

  • Possible evidence of superheating that might have triggered some of the high vesicularity internally in the lava flows and along contacts between magmas of different temperature/texture that were incompletely mixed prior to eruption;

  • Brown Cr spinel and altered olivine that have a ubiquitous and abundant presence in all extrusive rocks, whereby some rocks of Units IV and V carry enough olivine to be classified (originally) as picrite.;

  • Occurrence of ophimottle textures throughout Unit IV involving coarse-grained ophitic intergrowths of plagioclase and clinopyroxene set as loose networks within a finer grained matrix; and

  • Possibility of high-temperature alteration affecting all lithologies and resulting in the pervasive deep orange to red-brown staining in Units III and IV, as indicated by the presence of abundant secondary hematite and complex oxy-exsolution of titanomagnetite in Unit IV, which in turn contrasts with the chloritic nonoxidative alteration and the disappearance of titanomagnetite in Unit V.

Volcaniclastics and oligomict volcanogenic conglomerate (stratigraphic Subunit IIIb, lithologic Unit 3)

In thin section, both a large igneous clast (Fig. F21A, F21B) and a smaller clast of the conglomerate of Subunit IIIb (Fig. F21C, F21D) have abundant small plagioclase crystals arranged in a typical trachytic texture. Both clasts have abundant titanomagnetite, which is intact despite oxidative alteration of material between plagioclases (evident in the reddish background tint caused by internal reflections when using reflected light in Fig. F21B, F21D). The approximate modal abundance of the feldspar is 60%–70%, based on reflectivity. Different portions of the clasts have abruptly different plagioclase modes, sizes, and extent of crystallinity (Fig. F21A), and the plagioclase crystals flatten and turn to parallel vesicle margins where they concentrate to form narrow plagioclase-plated rims. The highly reflective titanomagnetite is skeletal in both samples and is as large as or larger than the plagioclase crystals, an attribute not typical of basalt, but rather of felsic differentiated lavas. The basaltic clasts are aphyric and have no residual clinopyroxene in between the small plagioclases that survived alteration. The large clast has several elongate vesicles as long as 1 cm, whereas the smaller clast in the conglomerate is nonvesicular.

Given all the basalt drilled at six sites during Ocean Drilling Program (ODP) Leg 198 and Expedition 324, these conglomerate clasts provide the only evidence for an extended range of differentiates at Shatsky Rise. The following brief inventory was made of all the rounded and irregular clasts, chips, and shards up to 1 cm in size within the conglomerate of Subunit IIIb, which are cemented together by calcite:

  • Several rounded clasts of ferroandesite, as described above.

  • Angular palagonitized granular hyaloclastite and hyaloclastic sandstone.

  • Some rounded and some angular lithic plagioclase-phyric basalt.

  • Lithic olivine-spinel-plagioclase phyric basalt, with altered olivine.

  • Angular and broken plagioclase phenocrysts and glomerocrysts.

  • Rounded lithic grains of mudstone.

  • Vesicular basalt cemented by quartz.

Palagonite and oxidation of most basalt and ferroandesite clasts in the conglomerate formed in situ and accompanied the rather similar alteration of immediately subjacent rocks.

Calcite and amygdule fillings in top of volcanic basement (stratigraphic Subunit IVa, lithologic Unit 4)

Thin Section 206 (Sample 324-U1349A-7R-1, 112–115 cm), taken ~20 cm below the top of volcanic basement in Unit IV, has a 1 cm wide crack and some large amygdules that are filled with calcite-cemented bioclastic material containing fossil fragments possibly derived from echinoderms and bryozoans. Some of the amygdules are also partly filled with palagonitized hyaloclastic sand, now cemented with calcite. The sand consists of former angular glass shards now entirely replaced by orange (partly fibrous) palagonite (Fig. F21E). About 10% of the material is dark red to brown angular fragments; these formerly had quenched spherulitic textures like those in zones of pillow basalts just adjacent to glass margins. A few more coarsely crystalline fragments (Fig. F21F) may be remnants of the deeper, more crystalline portions of pillows, or else fragments of fine-grained subaerial lava flows. The interiors of all basaltic fragments are nonvesicular, although some curving exterior outlines may have formed as bubble walls.

All of these sedimentary rocks and the extraordinary vesicularity of the topmost basalts in Unit IV indicate that eruptions occurred at, or near to, wave base. An ocean island must have been nearby in order to provide such a lithologically diverse volcanic detritus to a shallow offshore depositional environment that also supported a large calcareous macrofaunal assemblage. That assemblage, perhaps transported in alongshore currents, encountered and incorporated hyaloclastic material derived by eruption and quenching to volcanic glass in aquagene conditions. This mix of material covered the topmost basalt flow of Unit IV, after which the whole assemblage was soon altered and cemented.

Vesicular upper lava succession (stratigraphic Subunit IVa, lithologic Units 4–12)

Subunit IVa comprises eight lava inflation units defined primarily by highly vesicular flow tops with more massive interiors. In Core 324-U1349A-7R, six of the units have an average curated thickness of ~0.7 m. The more massive interiors have intervals with somewhat lower vesicularity and distinct variations in color from darker red to lighter brownish red that appear both as splotches or as streaks or bands trending diagonal to the core. The darker bands are usually narrower and slightly less vesicular than surrounding rock and typically outline lighter colored areas with contrasting sizes and spacings of the vesicles. These inflation units seem to have extruded as "lava foam" and may have become multicolored by differential oxidization upon eruption. The foam probably spilled over onto itself and tumbled around, a process that mixed together the bands with different vesicularity and color.

One thin section 2.31 m below the top of basement exhibits textural variability indicating very rapid cooling (Thin Section 209; Sample 324-U1349A-7R-3, 39–43 cm). The photos in Figure F22 show increasing crystallinity and development of crystal morphologies within ~2 cm of the edge of the thin section. Although the rock in Figure F22A is highly oxidized and vesicles are filled with calcite, the remnants of coalesced spherulites can still be identified. The few plagioclase crystals in the thin section are very tiny. Across the thin section, a successively larger size and higher abundance of clumps of plagioclase crystals can be seen (Fig. F22B–F22D). The rock also contains ~1%–3% olivine phenocrysts and small, intergrown crystals of Cr spinel. A question is whether it represents the chilled margin of a flow or pillow that erupted underwater. Quenched rims with (altered) glass are not evident in the cores of Hole U1349A, and it is possible that both vesiculation and intense alteration obliterated any glassy rims that were formed. Even so, the spherulitic margin shown in Figure F22A and the presence of hyaloclastite in cracks and vesicles require very high rates of cooling for crystallization to have been nearly or entirely suppressed.

Olivine phenocrysts in Subunit IVa are invariably altered to fibrous, moderately birefringent clays and red Fe oxyhydroxides (Fig. F23A, F23B). The red color is often most intense along crystal margins, a characteristic that in older studies (e.g., Baker and Haggerty, 1967) allowed it to be designated as iddingsite. Spinel is unaltered, but its dark amber brown matches the colors in altered surroundings. Each thin section contains a dozen or more grains and in some cases several grains per olivine crystal. The spinels retain cubic or octahedral shapes and usually occur associated with or within olivine (Fig. F23C, F23D) and often in clusters, but they also are isolated in the groundmass. Some crystals as large as 100 µm in diameter.

Vesicular lower lava succession (Subunit IVc, lithologic Units 14–31)

Externally, the rocks of Subunit IVc appear to be a continuation of the variegated and foamy lavas of the upper lava succession of Subunit IVa. However, dark bands and streaks of relatively fine grained and less vesicular rock are prominent in almost every section of core. These rocks appear to be mixtures of physically dissimilar lava types. Throughout Subunit VIc, the criteria for recognition of cooling or inflation units were a combination of high vesiculation and occurrence of a deep and pervasive red alteration, the combined observations giving the appearance of deeply weathered scoriaceous flow tops. This alteration presented a severe problem to identification of minerals in hand specimen. Outlines of olivine phenocrysts are fairly obvious, but many flow interiors in which acicular plagioclase could be discerned also contained a myriad of small and obviously altered bright red specks, which thin section examination revealed to be Fe oxyhydroxide minerals often coated with hematite (see below for details).

Pervasive alteration of olivine and the survival of spinel

Thin sections show that almost everything but plagioclase (present as needles in the groundmass) is altered to a combination of clays and Fe oxyhydroxides. In reflected light, the percentage of reflective silicate minerals is usually small (10%–40%), and nonreflective portions exhibit brown or red internal reflections (Fig. F23C, F23D), indicating the presence of abundant Fe oxyhydroxides in clays. In Figure F23C, plagioclase surrounding an altered olivine phenocryst is slightly darker than surviving interstitial clinopyroxene crystals. Elongate interiors of former glassy and crystalline material are now altered to dark brown and nonreflective secondary minerals. Only spinel survives unaltered.

Ophimottled textures

Many of the rocks of Subunit IVc are divided almost equally between finer grained and coarser grained groundmass domains, with the latter being distributed in loose networks across the thin sections. Whereas many rocks of previous sites drilled during Expedition 324 contain networks of intergrown plagioclase and clinopyroxene, in the coarser grained portions of these rocks the two are intergrown ophitically, in a fashion similar to that found in the centers of very massive flows, dikes, or gabbros. However, the domains of these intergrowths are visually surrounded by a dense speckling of Fe oxyhydroxides in the surrounding matrix of acicular plagioclase. This texture is termed "ophimottled," and the proportions of mottles clearly differ both within and between thin sections. For instance, in Figure F24A their distribution is defined by a difference in size and density of Fe oxyhydroxide grains to the left and right of a curving surface (delineated in part by the yellowish alteration). By contrast, the mottles in Figure F24D and F24E are larger and coarser grained.

Comparisons of individual mottles (Fig. F24) in transmitted and cross-polarized light are shown in Figure F25A–F25D, and between cross-polarized and reflected light in Figure F25E and F25F. Individual pyroxene crystals are 0.5–2 mm, and the mottles consist of one or more of such crystals, which (partly) enclose narrow, acicular, elongate, and often stellate arrangements of plagioclase crystals. Figure F25E and F25F shows a mantling of smaller plagioclase crystals that are partially intergrown, with the central clinopyroxene and later-formed clinopyroxene crystals arranged in a rim that contains grains of titanomagnetite. It is important to note that the mottles observed in thin section lack any faceted crystal outline and so are not strictly phenocrysts or glomerocrysts.

These plagioclase-clinopyroxene mottles are reminiscent of the intergrown gabbroic clots recovered with interstitial glass from the Juan de Fuca Ridge (Dixon et al., 1986). Their grain size and textures are also similar to those in some portions of shallow gabbros recovered near the top of the gabbroic section of ocean crust exposed at Hess Deep, eastern equatorial Pacific (e.g., Natland and Dick, 1996). Accordingly, the mottles in Hole U1349A probably represent crystallization at shallow levels within a gabbroic mass beneath Ori Massif; however, this crystallization did not proceed to completion; instead, the network of interlocking plagioclase-clinopyroxene clumps in the partially crystallized magma was disrupted by a late magma injection and then carried with this to the surface. The titanomagnetite in Figure F25E and F25F represents a fairly advanced stage in the basalt differentiation, but the mottles were raised to the surface upon eruption by a much more primitive magma than the liquids that produced these complex mottles. In the final host liquid, both olivine and spinel were stable.

Alteration of titanomagnetite

Some of the titanomagnetite in the ophimottles is skeletal (Fig. F26A, F26B), which means that portions of the crystals enclosed either melt or melt inclusions at the time the clots were incorporated into the ascending liquids. After quenching at eruption, however, all traces of that melt were erased by alteration, and the titanomagnetite itself underwent oxy-exsolution that may have been induced by the introduction of hot circulating fluids (Fig. F26B–F26D). Alternatively, these titanomagnetite crystals may have experienced an initial oxy-exsolution into solid solutions of ilmenite and magnetite before they were incorporated into the crystallizing lava. If the latter, the titanomagnetite crystals were well below solidus temperatures (but above ~600°C) at the time they were engulfed, and some of the features now shown in the minerals may have been inherited from these deeper processes.

Exsolution features can also be discerned in both the coarser oxide mineral grains of the mottles and the much smaller, skeletal, former titanomagnetite of the rest of the rock. The oxides in the mottled intergrowths and skeletal groundmass titanomagnetite therefore have similar features. In some larger crystals of the mottles, the oxy-exsolution produced a criss-crossing grid of four different oxide minerals in various pastel shades, as seen in Figure F26D. In some large oxide grains located at the edges of mottles (but otherwise in contact with the groundmass), these exsolution minerals appear to have been subject to further alteration and oxidation within the same thin section; in these instances, the crystals are more obviously gray, white, or pinkish gray in tone, and the crystals themselves having "ragged" or diffuse edges (Fig. F26D) within which internal reflections are bright red. The presence of four different oxide minerals in those examples may indicate that the grain previously was a gridwork of two minerals, these being ilmenite and magnetite solid solutions, and that these in turn underwent exsolution upon further oxidation: the possible oxide phases are ferrirutile, rutile, titanohematite, and pseudobrookite, within the framework of titanomagnetite oxidation Stages C4–C7 of Haggerty (1991). The presence of pseudobrookite would indicate that the exsolution temperatures reached 585°C (Lindsley, 1965).

Pillow basalt almost never contains titanomagnetite modified by deuteric alteration. This is because titanomagnetite in pillow lavas is usually tiny because it forms at very high cooling rates and because deuteric alteration typically only occurs during slow cooling of fairly thick lava flows or dikes. In Hole U1349A these complex oxy-exsolution features are very common, and exsolution lamellae occur even in the fine-grained rocks of Subunit IIIa (Fig. F27A, F27B) and in some very thin flow units interbedded with sedimentary rock. Additional exsolution and partial dissolution of titanomagnetite occurs in rock recovered 20 m deeper (Fig. F27C). Below this, at even deeper core intervals, skeletal titanomagnetite can have a remarkable array of exsolution features (Fig. F27D). In the deepest rocks of Subunit IVc (in Core 324-U1349A-13R) titanomagnetite is extensively dissolved (Fig. F27E) and narrow residual ilmenite lamellae form most of the residual framework of the oxide minerals. Finally, in the rocks of Unit V, titanomagnetite has vanished completely, which is evidenced by the rocks having low magnetic susceptibility values (see "Physical properties").

Development of hematite

Importantly, the brick-red color of basalt of Unit IV results from the presence of abundant Fe oxyhydroxides and hematite in the groundmass and not from iddingsite in altered olivine. These oxide minerals are closely spaced and fairly well crystallized (Fig. F28A) and are composites of both minerals, with the hematite being the most reflective (Fig. F28B). It forms thin veinlets that trace across both olivine and spinel and ringlets around deep red Fe oxyhydroxide crystals that lie between plagioclases (Fig. F28C). These oxides partly replace clinopyroxene (Fig. F28D), patches of intersertal glassy material, and the exterior portions of original titanomagnetite (Fig. F28E).

Vesicles and the consequences of magma mixing

In many volcanic circumstances, rocks with up to 75% vesicles (Fig. F29) are most unusual. Primitive olivine tholeiite and picrite that erupt on spreading ridges are not known to be volatile rich (Dixon, 1995; Asimow et al., 2004). Therefore, in very shallow water, such a huge vesicle percentage is unlikely to have been generated by simple magmatic degassing (Fig. F19). Not even Hawaiian tholeiite, which is more enriched than MORB, produces such a highly vesicular product, either in subaerial or submarine lava flows. However, it seems that the primitive basaltic rocks of Hole U1349A were practically foam when they erupted.

However, not all rocks of Unit IV have high vesicle abundances; for instance, the rocks with clear evidence for internal lava flow magma mixing show bands of significantly reduced vesicularity. The least vesicular part of the thin section in Figure F30 is the fine-grained diagonal nonophimottled domain. This less vesicular domain may have been the hotter of two constituent magmas, or it might have had a slightly different viscosity, yet it crystallized at the same cooling rate to a finer grain size. A similar contact between two different lithologies traverses vertically for a portion of the core, as shown in a thin section scan (Fig. F31A). In this figure, the highly vesiculated magma type on the left is more crystalline, and its altered mafic minerals are gray-green, standing out against plagioclase. The vesicles are infilled calcite cement and amount to nearly 75% of that part of the thin section. The orange rock on the right is more finely crystalline and contains only 10% vesicles, whereas the zone of magma mingling between the two has very few vesicles. Examination of the orange material at higher magnification reveals a sharp change in its crystallinity with distance away from the region of mingling. At the contact, the rock is extremely fine grained and spherulitic (Fig. F31B) but becomes more crystalline away from this contact, with better-formed plagioclase and (now mostly altered) dendritic pyroxene (Fig. F31C, F31D). Dark spots in the orange material are altered olivine that has an interior network of trellised hematite (Fig. F31E). This likely resulted from high-temperature oxidative alteration of the olivine, first exsolving the low-Ti magnetite, effectively taking all the iron in the magmatic olivine and leaving almost pure forsterite between the trellis work. Further oxidation transformed the magnetite to hematite (Haggerty and Baker, 1967).

These textures show that the orange magma type, carrying many olivine phenocrysts, was likely hotter and then chilled or quenched against the coarser grained gray magma type. The temperature difference between the two magma types might have been as high as 100°C. This magma mingling may have thus caused the host rock to "boil" along the contact with the higher temperature magma injection. Depending on the (yet undetermined) scale of such a superheating process, the locally released volatiles may have coursed through the lavas, producing a magma froth. After mixing with seawater, high abundances of these released volatiles could represent a possible agent that caused the oxy-exsolution in titanomagnetite at high temperatures during the eruption process, as well as a second source of oxygen to produce the brownish red rocks in Unit IV.

Flow breccia sequence (stratigraphic Unit V, lithologic Unit 32)

Unit V consists of basaltic flow breccias that formed under water. A number of breccia fragments were quenched to glass, but this is now totally altered (Fig. F32). Nevertheless, phenocrysts and tabular plagioclases can still be distinguished from the noncrystalline surroundings. In larger clasts, textures change in a similar fashion to those at the margins of pillow lava (i.e., to zones of isolated and coalesced spherulites; Figs. F32B, F33B, F33C) and then to sheafs and bundles of acicular plagioclase in an altered darker microcrystalline matrix (Figs. F32C, F32D, F33D). All of this material is now completely replaced by clays. No titanomagnetite can be seen, although some very tiny black crystals (too small to render a good surface for polishing) lie within the thickness of the thin section and may be residual ilmenite or secondary low-Ti magnetite, the vestigial remnants of alteration of titanomagnetite and clinopyroxene.

The original rocks may not have been identical in composition. Thin Section 258 (Sample 324-U1349A-16R-2, 7–14 cm) has a suggestion in its crystal morphologies of division into former plagioclase-clinopyroxene aggregates, perhaps mottles, that are surrounded by darker swaths of finer grained spherulitic material (Fig. F32D), whereas similar relict mottling is not evident in Thin Section 262 (Sample 324-U1349A-16R-6, 71–73 cm; Fig. F33). Spinel also occurs in the altered glass of the latter thin section (Fig. F33A) and every other more crystalline portion of the rock (Fig. F33B–F33E). Hopper crystallites of altered olivine occur in spherulitic portions of the thin section (Fig. F32B), and a number of peculiar radiating circular sheafs of a fibrous mineral with moderately high birefringence (Fig. F32F) could be incipient bundles or rosettes of clinopyroxene.

Olivine phenocrysts probably are ~5% of any thin section in Unit V, but the olivine is invariably pseudomorphed by calcite (Fig. F34A) or clays and Fe oxyhydroxides (Fig. F34B). In the rocks of Unit V, hematite is all but gone as a secondary replacement in olivine (Fig. F34C–F34E). Only Cr spinel survived most of the alteration in Unit V, although reticulate alteration and pitting does occur (Fig. F35A, F35B). The spinel is usually brown, but where unaltered and relatively thin it has a slightly greenish tint (Fig. F35C), suggesting a rather aluminous composition. Otherwise, the mineral occurs both in olivine phenocrysts and scattered through spherulitic groundmasses amid the acicular plagioclases and buttonlike pyroxene rosettes (Fig. F35D, F35E).

One thin section has a peculiar large vermicular Cr spinel phenocryst with an overall faceted outline (Fig. F36A) that must represent a type of skeletal crystal morphology. In reflected light, the spinel exhibits its usual triangular octahedral morphology (Fig. F36B). A similar crystal in basalt from the Mid-Atlantic Ridge was described by Eisenach (1979). The reflectivity is much lower than that of titanomagnetite, which in any case is absent from this particular thin section because of alteration. High magnification shows that the crystals are indeed brown with a greenish tint, like other spinel in Unit V, and not black, the color of titanomagnetite (Fig. F36C).

Preliminary assessment

The basalt in Hole U1349A is the most primitive rock recovered during Expedition 324. However, further revealing its primary igneous chemistry will require disentangling the profound effects of alteration seen in the rocks of stratigraphic Units IV and V. Aspects of its mineralogy, especially the compositions of pyroxenes and intergrown plagioclase in the ophimottled rock and of Cr spinel throughout, will aid in addressing this task. Perhaps the most pressing question involving Units IV and V concerns the formation of olivine-replacing hematite in the basalt matrix and the pervasive oxy-exsolution of most (if not all) of the titanomagnetite, which arguably occurred at temperatures of ~600°C. One possibility is that the basaltic magmas were initially oxidized and oxide minerals were transformed at high temperatures because of their encounter with juvenile and meteoric fluids accompanying the boiling associated with the magma injection and mixing. Afterward, alteration would have then taken a more usual course, becoming mainly an overprint produced by normal interaction with fluids modified from seawater during hydrothermal or low-temperature circulation. The dusky green color, the disappearance of titanomagnetite, and loss of the magnetic susceptibility signal in Unit V suggest that temperatures during this second-stage bulk alteration were still elevated. Chemical analyses indicate that the process of alteration significantly leached K2O and added MgO to these rocks, whereas the opposite effects occurred in the rocks of Unit IV.