IODP Proceedings    Volume contents     Search

doi:10.2204/iodp.proc.309312.103.2006

Expedition 312

Igneous petrology

Hole 1256D was deepened from ~1255 to ~1507 mbsf during Expedition 312. The first core, a ghost core containing loose material accumulated at the bottom of the hole prior to and during reentry operations, was Core 312-1256D-171G, which continues the numerical sequence from Expedition 309 (Core 309-1256D-170R). Cores 312-1256D-172R through 234R were drilled during Expedition 312. Recovery varied according to lithology, with low returns from sheeted dike intervals and generally high returns from gabbro intervals. Average overall recovery was 18.8%.

As the basis for our core descriptions, we defined lithologic units (Fig. F198) based on criteria observable in hand specimen, including mineralogy, texture, and grain size (see the “Methods” chapter). For Expedition 312, which recovered only intrusive rocks, the principal objective of the unit designations is to ensure that individual cooling and/or intrusive units are recognized and retrievable from the databases. Unit numbers continue the sequence begun during Leg 206 (Wilson, Teagle, Acton, et al., 2003) and continued through Expedition 309. The first unit from Expedition 312 (Unit 1256D-66) appears to be a continuation of Unit 1256D-65, the last unit of Expedition 309. Thirty new units were identified during Expedition 312 (Units 1256D-66 through 95) (Table T35). In a few cases, subunits were defined to record the presence of small igneous dikes and other small but important lithologic features (see the “Methods” chapter).

Coring in Hole 1256D resumed in sparsely plagioclase-phyric to aphyric fine-grained massive basalt of the sheeted dike complex. Between ~1255 and ~1407 mbsf, we continued to recover predominantly aphyric fine-grained to microcrystalline basalts with uncommon cryptocrystalline intervals, in some cases grading inward from presumed dike margins. In several cases, near-complete cooling units graded from cryptocrystalline margins to fine-grained centers were recovered. All basalts from the sheeted dike complex have been substantially altered, and both the intensity and grade of alteration increase downhole (see “Alteration”). The uppermost units in Expedition 312 cores are only partially altered to greenschist-grade mineral assemblages, whereas those near the base of the sheeted dike complex (>1348 mbsf) are completely altered to amphibolite-grade assemblages with localized higher grade domains, characterized by secondary pyroxene, that approach granulite conditions. These domains typically exhibit granoblastic textures, and their onset defines the “granoblastic dike” sublithology of the sheeted dike complex. At 1406.6 mbsf in Hole 1256D, we encountered a gabbro dike, marking the end of the sheeted dike complex and the beginning of the “plutonic section.”

The plutonic section of Hole 1256D includes two gabbroic intervals: Gabbro 1 (1406.6–1458.9 mbsf; Units 1256D-81 through 89) and Gabbro 2 (1483.1–1507.1 mbsf; Units 1256D-91 through 95). Gabbro 2 consists of three separate igneous intervals: an upper, medium-grained orthopyroxene-bearing gabbro that grades to gabbronorite near its margins (Units 1256D-91 through 93), a fine-grained gabbronorite of which the origin is unclear (Unit 1256D-94), and a later basaltic crosscutting dike (Unit 1256D-95).

The medium-grained gabbroic bodies are intrusive and have produced high-temperature contact metamorphic effects in the host metabasalts of the sheeted dike complex. Gabbro 1 and Gabbro 2 are separated by a dike screen (1458.9–1483.1 mbsf; Unit 1256D-90), a metabasalt interval with a well-developed “granoblastic” metamorphic overprint.

In the following section, we present macroscopic observations, petrographic observations, and descriptions of lithologic units.

Overview of macroscopic observations

Sheeted dike complex (1060.9–1406.6 mbsf; Units 1256D-66 through 80)

The sheeted dike complex begins at 1061 mbsf, below the last occurrence of thin sheet flows or other features associated with lava flows, and continues uninterrupted to 1406.6 mbsf, where it is first intruded by plutonic dikes. We divided the Expedition 312 portion of this interval into 15 lithologic units (Units 1256D-66 through 80; 1255–1406 mbsf), based primarily on abrupt changes in texture and/or grain size. In the lower part of this sheeted dike interval, below 1348.8 mbsf, the onset of a high-temperature metamorphic paragenesis characterized by locally developed granoblastic textures defines a subzone, the granoblastic sheeted dikes. A second sheeted dike interval, the dike screen, extends from 1458.9 to 1468.6 mbsf, separating the first plutonic interval (Gabbro 1) from the second plutonic interval (Gabbro 2). The dike screen is described in “Plutonic section (1406–1507 mbsf; Units 1256D-81 through 95).”

Because the trajectory of the drill hole is subparallel to the dip of the dikes, it is difficult to relate lithologic units and other features in the core to individual discrete dikes. Recovery of the sheeted dike complex was <12%, making reconstruction of the dike sequence difficult. Whenever oriented igneous contacts were recovered in the sheeted dike complex, they are subvertical and associated with distinctive cryptocrystalline chilled margins (Figs. F199, F200). Some of these contacts coincide with lithologic unit boundaries, but some do not. These internal boundaries may represent instances where the drill hole intersected an irregularly shaped dike margin and then returned to the interior of the same dike.

Three intrusive contacts were recovered from unit boundaries within the sheeted dike complex during Expedition 312 (Units 1256D-68/69, 75/76, and 80/81). In addition, cryptocrystalline material is present at the Unit 1256D-77/78 boundary, and Unit 1256D-71 is a thin unit that is entirely cryptocrystalline with an internal, brecciated contact, most likely representing the contact between Units 1256D-70 and 72. An additional three units (Units 1256D-70, 71, and 73) have internal contacts. Individual dikes may also be identified on the basis of downhole grain size gradations from finer to coarser and back to finer. Six such cooling units are recognized in Expedition 312 cores. Taken together, these lithologic data indicate at least 9 individual dikes among the 12 sheeted dike units defined during Expedition 312 (Fig. F201). This estimate relies upon the assumption that individual dikes are not repeated downhole, either by faulting or because of the meandering of an individual intrusion.

Geochemical data can also be used to assess the number of dikes present in Expedition 312 cores. Lithologic evidence can sometimes be used to confidently relate pairs of geochemical samples to a single dike. One such case occurs for the two geochemical samples taken from Unit 1256D-69. These sample pairs tend to have similar compositions, particularly for trace element ratios that are insensitive to in situ differentiation processes (e.g., Zr/Y). The implication of this observation is that geochemical variation within individual dikes is limited. However, of the nine dikes defined on lithologic evidence, at least two show internal variation greater than that expected for a single intrusion: Dikes 1 and 9 appear to contain at least three distinct compositions (Fig. F201). Therefore, if individual dikes are not compositionally zoned and do not repeat within the hole, a combination of lithologic and geochemical data indicates that a minimum of 13 dikes were sampled in Expedition 312 cores.

The sheeted dikes are, for the most part, uniform, aphyric massive basalts. In most respects, visual and thin section descriptions from Expedition 309 closely match those from Expedition 312. There are, however, significant changes in grain size and phenocryst abundance at 1255 mbsf, where coring resumed for Expedition 312 (Fig. F198). At 1255 mbsf, the predominant grain size changes downhole from microcrystalline to fine grained. We believe that this change is real, not an artifact of the personnel change, because we were very careful to calibrate our Expedition 312 grain size estimates by creating visual calibration cards from actual Expedition 309 core pieces (see the “Methods” chapter). Below 1340 mbsf, there is a second shift in the predominant grain size, which becomes smaller on average and more variable. Grain size terms in this report are defined in the “Methods” chapter.

Leucocratic lithologies

The only leucocratic rock recovered in core from the sheeted dike complex is a 20 mm wide trondhjemite dike (Unit 1256D-80b) (Sample 312-1256D-212R-1, 29–32 cm; 1404.4 mbsf) from ~2 m above the plutonic complex. However, leucocratic igneous lithologies are relatively abundant (~1%) in the coarse sand-size fraction of the material recovered in junk baskets during hole clearing operations. These samples are typically fine grained with a variety of primary intergranular textural types. They consist of quartz, highly altered feldspar, and actinolitic hornblende, the latter presumably replacing a primary mafic phase. Deformation lamellae, wavy extinction, and partial to complete clastic recrystallization indicate significant deformation (see “Unit 1256D-79: mixed aphyric fine-grained basalts and leucocratic fragments”).

Plutonic section (1406–1507 mbsf; Units 1256D-81 through 95)

At least two gabbro bodies intrude the sheeted dike complex, followed some time later by the small crosscutting basalt intrusion of Unit 1256D-95 (Fig. F202).

Upper contact of the plutonic section

The first occurrences of equigranular, generally medium-grained “plutonic” rocks in Hole 1256D are two small gabbro dikes that intrude Unit 1256D-80 (interval 312-1256D-213R-1 [Pieces 12 and 13, 48–63 cm]; 5 and 40 mm, respectively). These dikes define the upper boundary of Unit 1256D-81 and represent the top of the plutonic section (Fig. F203). As described previously, a 5–20 mm wide microcrystalline trondhjemite magmatic dike (Unit 1256D-80b) also intrudes basalt from Unit 1256D-80a at ~1404.3 mbsf. This thin leucocratic dike is not formally included in the plutonic section.

Gabbro 1

With the exception of the thin uppermost units (Units 1256D-81 and 82), Gabbro 1 is mineralogically and texturally heterogeneous. It is divided into seven units (Units 1256D-83 through 89), primarily on the basis of relatively abrupt variations in dominant texture and/or grain size, although both texture and grain size vary locally on a centimeter scale as well. Gabbro 1 gabbros are commonly oxide bearing, and oxide abundance decreases irregularly downhole (Fig. F204). Olivine is present in significant amounts only in the lower two units (Units 1256D-88 and 89).

An important feature that these seven units have in common is the development of a patchy texture defined in hand specimen by the contrast between subspherical dark domains and an intervening network of lighter colored material (Fig. F205).

A highly unusual rock type is present near the top of Gabbro 1. Unit 1256D-82 is a thin (<1 m) leucocratic dike that intrudes the gabbro. Its mineralogy is that of an oxide-rich tonalite, but it is compositionally equivalent to an evolved oceanic FeTi basalt (see “Unit 1256D-82: quartz-rich oxide diorite”). We use the name “quartz-rich oxide diorite” for this lithology.

Dike screen

The dike screen (Unit 1256D-90) is an interval of sheeted dikes trapped between the younger intrusions of Gabbro 1 and Gabbro 2. It consists of predominantly fine-grained metabasalt with a contact metamorphic overprint characterized by granular secondary clinopyroxene and lesser orthopyroxene obscuring the original igneous texture. The dike screen is cut by a number of small quartz gabbro and tonalite dikes, varying in thickness (1–10 cm), grain size, and composition.

Gabbro 2

The upper, orthopyroxene-bearing gabbro interval of Gabbro 2 is characterized by an absence of fresh olivine, high but variable orthopyroxene contents (5%–25%), and considerable local heterogeneity. Oxides are present throughout, ranging to 5% and rarely higher. Oxide abundance generally diminishes downhole (Fig. F204). The predominant rock type is orthopyroxene-bearing gabbro with gabbronorite in the marginal units (Units 1256D-91 and 93). At the upper margin of this zone, gabbronorite invasively intrudes the dike screen, isolating and detaching centimeter-sized blocks of metabasalt. Similar detached blocks of metabasalt are also present close to the lower contact where the boundary of the intrusion contains several fragments of pinkish orthopyroxene-rich metabasalt. The fragments and the main body of Unit 1256D-94 are orthopyroxene-rich, recrystallized cryptocrystalline to fine-grained basaltic rocks.

Below the medium-grained gabbro of Gabbro 2 is a fine-grained “gabbronorite” of uncertain origin (Unit 1256D-94). It is extensively recrystallized with locally well-developed granoblastic textures, although relict igneous intergranular textures are preserved in some places, especially away from the upper boundary. There is some ambiguity as to whether this unit is a metabasalt derived from younger sheeted dikes or a fine-grained intrusive gabbronorite (see “Unit 1256D-94: fine-grained gabbronorite or metabasalt”).

Beginning at 1502.6 mbsf, Unit 1256D-95 is a relatively light gray cryptocrystalline basalt that is distinct from other Expedition 312 basalts by abundant pink clinopyroxene (titanaugite?) and a low primary oxide content. It is also distinct from other basaltic rocks of the plutonic section in that it exhibits greenschist-facies alteration minerals similar to shallower dikes from above Unit 1256D-70 (~1275 mbsf). We interpret this unit as a crosscutting, relatively enriched intrusive that intruded after significant cooling had occurred.

Overview of petrographic observations

Sheeted dike complex
Upper dikes

During Expedition 312, we recovered predominantly fine-grained massive basalts that grade locally to centimeter-scale microcrystalline domains and less commonly to plumose and spherulitic quench textures in cryptocrystalline chilled margins. Deeper in the hole, dominantly microcrystalline intervals become more common. Locally coarser patches are also common (e.g., Fig. F206).

Typical textures are hypocrystalline intergranular to intersertal with mesostasis completely altered, mainly to chlorite in the upper part of the Expedition 312 core, changing progressively to fibrous brown cryptocrystalline material (becoming dominant in Unit 1256D-72, below 1277.8 mbsf) and, at greater depth, to actinolite (beginning in Unit 1256D-74, at 1305.5 mbsf) (see “Alteration”). Plagioclase typically forms a textural framework that is predominantly randomly oriented (intergranular), commonly grading to variolitic or radiate (characterized by radiating prisms). Plagioclase and clinopyroxene are most commonly moderately altered (see “Alteration”). Fe-Ti oxides are abundant throughout the section as relatively large primary subhedral to euhedral grains and abundant tiny anhedral grains within fibrous brown and actinolitic alteration areas.

Clinopyroxene and plagioclase occur uncommonly as small phenocrysts and rarely together as small 2–3 mm sized glomerocrysts. Olivine occurs in rare glomerocrysts with clinopyroxene and plagioclase in the less altered samples, generally those from shallower depths.

The majority of basalt samples from Expedition 312 are aphyric, with only 4 of the 55 thin sections containing >1% phenocrysts (Fig. F198). This scarcity of phenocrysts contrasts with the variable and higher proportions of phenocrysts above 1255 mbsf, where >25% of the thin sections contain >1% phenocrysts. Although this change may, in part, result from the greater difficulty of distinguishing very small crystals present in incoming magmas from a coarsening fine-grained groundmass, examination of Expedition 309 thin sections and the general absence of phenocrysts in the finer grained rocks below 1340 mbsf suggest that this difference is also not simply a product of the change in personnel.

Granoblastic dikes

Basalts in the dikes between 1348.3 mbsf and the bottom of the sheeted dikes (1406.62 mbsf) (Sections 312-1256D-192R-1 through 213R-1 [Piece 13, 52 cm]) are characterized by patchy areas having distinct granoblastic textures that define a separate subdivision within the sheeted dike complex. The key feature of these rocks is the presence of localized centimeter-scale granoblastic patches with common secondary clinopyroxene (see “Alteration”). Secondary clinopyroxene occurs as small, equant, anhedral grains intergrown with recrystallized plagioclase (Fig. F207B, F207C). Secondary clinopyroxene in some cases contains blebs of secondary magnetite (e.g., Figs. F208B, F208C, F209A). Primary titanomagnetite is completely recrystallized to subequant grains of secondary magnetite and ilmenite. Small anhedral orthopyroxene is also present in some of these granoblastic patches and in some veins. The proportion of granoblastic patches varies within this interval (see “Alteration”). Outside the granoblastic patches, typical igneous textures, overprinted by hydrous alteration, are present. Secondary assemblages are amphibole rich, with primary clinopyroxene pseudomorphed by green amphibole containing disseminated tiny grains of secondary magnetite (Fig. F207B, lower left). Plagioclase is highly altered to secondary plagioclase, and titanomagnetite is recrystallized to magnetite and ilmenite with trace titanite.

In Unit 1256D-80, which is immediately above the first gabbro unit, basalt is highly to completely recrystallized (see “Alteration”). Common in these rocks are microcrystalline equigranular domains consisting of granular plagioclase, tabular anhedral green amphibole, granular anhedral colorless clinopyroxene, and larger subhedral oxide grains (e.g., Fig. F209A–F209C). Orthopyroxene occurs in two modes: around metamorphic veins and coexisting with clinopyroxene in some microgranular domains (e.g., Fig. F209C).

Mineralogy and textures in the patchy domains in the granoblastic dikes are consistent with high-temperature contact metamorphism, possibly under upper amphibolite- or granulite-facies conditions. The granoblastic dikes indicate a steep thermal gradient over ~50 m through the lower dikes above the underlying intrusive gabbros. The temperatures at which these assemblages formed and their specific reaction paths will be of considerable interest for later study.

Plutonic section
Gabbro 1

Thin section inspection shows that the dark subspherical domains comprise individual clinopyroxene oikocrysts mostly 5–10 mm in diameter (Fig. F210). They enclose relatively small unzoned prismatic plagioclase grains and are partially surrounded by hornblende reaction coronas. The lighter colored network domains are coarser grained, seriate, intergranular gabbro with strongly zoned plagioclase, abundant oxides, and symplectic quartz/​plagioclase intergrowths. This texture seems to require two magmatic “episodes.” During the first episode, a plagioclase framework in or near equilibrium with developing clinopyroxene oikocrysts is developed. The second episode introduces new magmatic liquid that disrupts and partially disaggregates the preexisting framework.

Quartz-rich oxide diorite consists of quartz (20%–25%), abundant (>20%) interstitial Fe-Ti oxides, hornblende, and plagioclase, both as subhedral tabular prisms with significant zoning and as granophyric intergrowths with quartz.

Gabbro 2

Gabbros in Gabbro 2 consist of medium-grained gabbro-gabbronorite (with disseminated oxides in some cases). Constituent minerals of the gabbros are plagioclase, clinopyroxene, orthopyroxene, and oxides. Minor amounts of altered olivine (up to 1%) also occur in some cases. These gabbros are characterized by the common appearance of orthopyroxene and scarce presence of olivine compared to gabbros in Gabbro 1. Inequigranular seriate to poikilitic textures are most common. Poikilitic textures are characterized by oikocrystic clinopyroxene and orthopyroxene that include euhedral small (0.05–0.4 mm in long axis) tabular plagioclase grains. Some medium-grained gabbronorites have textures that make them virtually indistinguishable from metamorphosed basaltic fragments. Lithologic contacts between fine-grained gabbronorites and medium-grained gabbros are commonly diffuse in thin section. Relatively coarse quartz-bearing lithologies (e.g., medium-grained oxide quartz-diorite; Sample 312-1256D-232R-2, 0–3 cm [Thin Section 114]) are also observed in some intervals. This lithology consists of clinopyroxene, plagioclase, amphiboles, quartz, and oxides and shows inequigranular seriate texture with heavy alteration.

Downhole metamorphic evolution

All basaltic samples recovered from the sheeted dike section of Hole 1256D during Expedition 309/312 have a distinct metamorphic overprint (see “Alteration”). There is a rapid downhole increase in both alteration intensity and metamorphic grade with the transition from partial replacement of primary clinopyroxene by dusty actinolitic amphibole to the appearance of secondary clinopyroxene occurring over <100 m (Fig. F211). Alteration of the underlying gabbros is significant, and in some samples the distinction between secondary and igneous processes is difficult to distinguish. In order to clearly document the changes in metamorphic overprint with depth in the sheeted dikes, we focused on the textural evolution of fine-grained samples with progressive metamorphic overprint. We define seven textural types, assigning qualitative rank numbers.

Intersertal texture is characterized by a framework of plagioclase laths, commonly associated with prismatic clinopyroxene, between which angular spaces are occupied by finer crystals, glass, or their alteration products. The framework plagioclase is commonly skeletal with elongate, cylindrical hollow cores; radiating clusters of plagioclase prisms and variolitic domains characterized by radiating clusters of finer grained plagioclase are common local variants. The defining characteristics of the seven textural types are given in Table T36. Representative thin section images of textural Types 2–7 are shown in Figure F212.

Development of secondary (metamorphic) pyroxene

The shallowest occurrences of significant secondary clinopyroxene are characterized by isolated single crystals of anhedral granular clinopyroxene forming a metamorphic paragenesis with actinolitic hornblende, albite-rich plagioclase, and magnetite (see “Alteration”). This paragenesis occurs as patchy areas of typical granoblastic texture (Fig. F208) interspersed through larger areas in which actinolitic assemblages replace primary mafic phases within an inherited, primary intersertal texture (Figs. F208A, F212E).

The apparent fragmentation of primary prismatic clinopyroxene into secondary anhedral grains begins with alteration of primary clinopyroxene to fibrous actinolite plus magnetite during initial hydrous alteration (Fig. F208A). As a consequence of reheating, relict clinopyroxene (Fig. F208B, obscured object on right) is converted to discrete clusters of small secondary anhedral clinopyroxene that still incorporate tiny oxide grains (Fig. F208B, left side). With increasing temperature, patches of isolated small anhedral to subhedral secondary pyroxene grains may form. At the highest metamorphic conditions within the sheeted dikes, an equigranular framework of anhedral clinopyroxene, plagioclase, amphibole, and oxides has formed (Fig. F208C, lower image). The loss of amphibole from these domains is a dehydration process, and the reasons for the patchiness are unclear. Some possibilities include (1) locally varying fluid compositions, (2) heterogeneities of the starting material (i.e., different textural precursors such as variolitic and intergranular domains within the same section), or (3) precursors related to irregular patches of low-grade groundmass alteration.

Figure F209A illustrates the proposed sequence, albeit within a different primary texture. Here, a relict clinopyroxene phenocryst contains magnetite inclusions within material having a distinct discoloration and cleavage along the rim of the crystal and along boundaries with an intergrown plagioclase crystal. These occurrences may reflect initial hydrothermal alteration of clinopyroxene to amphibole plus magnetite along grain boundaries. This initial alteration was followed by wholesale recrystallization of phenocrysts to secondary clinopyroxene, preserving secondary magnetite as inclusions. There is some residual amphibole in embayments along the lower edge of the phenocryst. A similar process may occur on a finer scale in the granoblastic patches. In rare instances, carbonate is also present in granoblastic patches (Fig. F209B; see “Unit 1256D-80: aphyric cryptocrystalline basalt” in the “Appendix”).

Shortly below the first occurrences of secondary clinopyroxene, small amounts of orthopyroxene also appear (Samples 312-1256D-194R-1, 36–37 cm, and 196R-1, 32–33 cm [Thin Sections 35 and 36]). In some places, orthopyroxene is present as small (<0.1 mm) prismatic crystals in the interstices between plagioclase laths. Elsewhere, small anhedral orthopyroxene grains coexist with secondary clinopyroxene.

Throughout the section, primary minerals are pervasively altered to amphibole-dominated assemblages. Higher grade domains with a clinopyroxene-orthopyroxene–dominated paragenesis heterogeneously occur irregularly throughout the section. These become more abundant downhole and predominant within and below the dike screen (below ~1400–1450 mbsf). Within this higher grade part of the section, orthopyroxene is abundant and variable in form. In places, it appears to have a magmatic origin. In others it is clearly metamorphic. This complexity emphasizes that these basal units represent an important interface where the distinction between magmatic and metamorphic processes is difficult.

Unit descriptions

In the following sections, we briefly describe the igneous units cored during Expedition 312. Detailed descriptions are provided in the “Appendix.”

Sheeted dike complex
Unit 1256D-66: sparsely plagioclase-phyric to aphyric fine-grained basalt

Basalts of this unit are fine grained with local coarser patches (Fig. F206). Interstitial glassy material is mainly altered to chlorite (Fig. F213A, F213B), and clinopyroxene is partially altered to brownish, fibrous, cryptocrystalline material. Clinopyroxene-plagioclase glomerocrysts, rarely with olivine, are scattered throughout (Fig. F213C, F213D). The metamorphic overprint in the intersertal textures corresponds to texture Type 3 (Table T36).

Unit 1256D-67: mixed aphanitic and fine-grained basalts

This unit comprises out-of-place, mixed cryptocrystalline and fine-grained lithologies from Units 1256D-66 and 68 disturbed during drilling.

Unit 1256D-68: aphyric fine-grained basalt

This unit is characterized by visible fine-grained acicular plagioclase. Alteration is similar to that of Unit 1256D-66, texture Type 3 (Figs. F214, F215).

Unit 1256D-69: aphyric cryptocrystalline to fine-grained basalt

The upper contact is intrusive into Unit 1256D-68 (Fig. F216), and this unit appears to be a complete cooling unit, grading downward from a cryptocrystalline upper chilled margin through a fine-grained center to a cryptocrystalline lower margin (Fig. F217A–F217C), but no lower contact was recovered. Sparse microphenocrysts of euhedral clinopyroxene and clinopyroxene-plagioclase clots are present within the cryptocrystalline marginal material (Fig. F216). Alteration is unchanged texture Type 3.

Unit 1256D-70: aphyric microcrystalline to fine-grained basalt

This unit is distinguished from Unit 1256D-69 by abundant acicular plagioclase (Figs. F218, F219). Clinopyroxene is more altered to brownish, dusty, cryptocrystalline material containing abundant, tiny, anhedral oxide grains. Interstitial glassy material is completely altered to chlorite plus oxide. The metamorphic overprint is intermediate between texture Types 3 and 4.

Unit 1256D-71: brecciated cryptocrystalline basalt

This short unit consists of several pieces of brecciated cryptocrystalline, aphyric basalt (Fig. F200) that contain an intrusive contact and complex internal textural changes (Fig. F220). It includes brecciated margin(s) of one or more basaltic dikes.

Unit 1256D-72: aphyric microcrystalline to fine-grained basalt

Grain size increases downward from microcrystalline to fine grained, reflecting a transition from dike margin to dike interior, but no contacts were recovered. Plagioclase and minor clinopyroxene microphenocrysts form clots in the finer grained upper part of the unit (Fig. F221). Alteration overprint is unchanged, between texture Types 3 and 4.

Unit 1256D-73: aphyric cryptocrystalline to fine-grained basalt

This unit consists of small broken pieces of microcrystalline basalt from an interval of poor recovery. A single cryptocrystalline basalt piece includes an intrusive contact (~1290.5 mbsf; interval 312-1256D-179R-1 [Pieces 1 and 2, 0–9 cm]), but unit boundary contacts were not recovered. Fibrous actinolite appears for the first time as an alteration product of clinopyroxene and former mesostasis. Fine, second-generation Fe-Ti oxide grains also become prominently dispersed in actinolite. These changes are characteristic of metamorphic texture Type 4 (Fig. F222).

Unit 1256D-74: aphyric cryptocrystalline to microcrystalline basalt

Finer in grain size than Unit 1256D-73, Unit 74 is a thin grain size–graded cooling unit. Fibrous actinolite and associated disseminated fine Fe-Ti oxides continue to be the dominant alteration products (texture Type 4) (Fig. F223).

Unit 1256D-75: aphyric microcrystalline to fine-grained basalt

Unit 1256D-75 is a single cooling unit defined by visible acicular plagioclase (Fig. F224A). The intensity of the alteration overprint is greater than in Unit 1256D-74 with higher abundances of fibrous actinolite and secondary Fe-Ti oxide, transitional between texture Types 4 and 5.

Unit 1256D-75b: aphyric cryptocrystalline basalt dike

Unit 1256D-75b is a narrow (~1 cm) kink-banded, cryptocrystalline basalt dike that intrudes the slightly coarser basalt of Unit 75 (Figs. F224B, F224C, F225).

Unit 1256D-76: aphyric cryptocrystalline to microcrystalline basalt

Basalt from Unit 1256D-76 is predominantly cryptocrystalline, grading downhole to microcrystalline. Primary textures are obscured by fibrous actinolite, corresponding to texture Type 5. Unit 1256D-76 is intruded by Unit 1256D-75 (Fig. F226).

Unit 1256D-77: aphyric fine-grained basalt

This single cooling unit of predominantly fine-grained basalt grades to almost medium grained near the unit center at 1334.5 mbsf and then to microcrystalline at the lower unit boundary near 1343.5 mbsf. Alteration generally corresponds to Type 5, as in Unit 1256D-76. Rare patches in which primary clinopyroxene appears to have been recrystallized into granular clusters represent the first downhole occurrence of secondary clinopyroxene.

Unit 1256D-78: aphyric microcrystalline to fine-grained basalt

Although they are lithologically identical, Unit 1256D-78 was defined to distinguish it from Unit 1256D-77, which is a single cooling unit. In fact, geochemical variation within Unit 1256D-78 suggests that it consists of two compositionally distinct dikes (Fig. F201). Unit 1256D-78 encompasses a marked downhole increase in the intensity and grade of alteration. Primary textural and structural features (Fig. F209A) are increasingly obscured by dense fibrous actinolite with secondary tiny Fe-Ti oxide. Brownish inclusion-free amphibole (hornblende?) occurs for the first time.

A key feature of this unit is the appearance of granoblastic patches dominated by secondary clinopyroxene (Fig. F207B, F207C), rarely accompanied by very small anhedral orthopyroxene, forming a high-temperature metamorphic paragenesis presumed to result from heating by gabbroic intrusions. The metamorphic overprint increases from texture Types 5 to 7 through this unit, indicating a remarkably steep thermal gradient.

Unit 1256D-79: mixed aphyric fine-grained basalt and leucocratic rock fragments

Unit 1256D-79 includes all the geological material collected in junk baskets during hole-cleaning operations. A single basalt sample appears to be from Unit 1256D-78, but the majority of pieces, including numerous fine-grained leucocratic rock fragments of varied lithology present in the gravel-sieve fraction, may be from shallower parts of the hole.

Unit 1256D-80: aphyric cryptocrystalline basalt

Unit 1256D-80 spans a wide interval of very low recovery throughout which the recovered samples are surprisingly uniform, partly reflecting extensive metamorphic recrystallization that has largely obscured the primary textures. Two small gabbroic dikes near 1406.6 mbsf define the top of the underlying plutonic section and the beginning of the uppermost gabbro unit.

Most Unit 1256D-80 basalts have completely recrystallized under amphibolite-facies conditions, forming a variety of heterogeneous domains. Microcrystalline equigranular domains include granular plagioclase, green actinolitic hornblende, granular clinopyroxene, and larger subhedral oxide grains (Fig. F209A, groundmass). Amphibole forms rare poikiloblastic clusters enclosing plagioclase and oxide. Secondary clinopyroxene-rich domains apparently recrystallized from actinolite that has previously replaced magmatic clinopyroxene (Fig. F209A). Some of these domains contain minor orthopyroxene (Fig. F209C). Patchy orthopyroxene is also present around some metamorphic veins, probably related to active or precursory hydrothermal alteration. The metamorphic overprint in Unit 1256D-88 approaches granulite-facies conditions and defines texture Type 7, the highest grade found in the sheeted dike complex.

Unit 1256D-80b: trondhjemite dike

Unit 1256D-80b is a 5–20 mm wide magmatic dike of microcrystalline trondhjemite (Figs. F227, F228).

Plutonic section
Gabbro 1: Units 1256D-81 through 89

The medium-grained gabbros of Gabbro 1 are divided into nine units (Units 1256D-81 through 89), primarily on the basis of textural and grain size changes. Most of Gabbro 1 is characterized by a patchy domain structure that varies in scale and relative proportions but is almost always present. Dark subspherical domains as large as 30 mm in diameter are dominated by anhedral poikilitic clinopyroxene enclosing unzoned prismatic plagioclase, a textural feature more typical of dolerites. Between the darker domains, a lighter colored network domain consists of coarser grained oxide-rich equigranular gabbro in which plagioclase is distinctly larger and strongly zoned. Primary Fe-Ti oxides are present throughout Gabbro 1, but their overall abundance decreases downhole. Olivine is present in significant amounts only in Units 1256D-88 and, especially, 1256D-89.

Unit 1256D-81: intermixed medium-grained oxide gabbro and basalt

Two very small gabbro dikes that intrude the metabasalt from Unit 1256D-80 at ~1406.6 mbsf (Fig. F203) define the beginning of the plutonic section and Unit 1256D-81 in Hole 1256D. The remainder of Unit 1256D-81 consists of mixed medium-grained gabbroic and basaltic rubble.

Unit 1256D-82: quartz-rich oxide diorite

Unit 1256D-82 is a narrow (<1 m) intrusion into Units 1256D-81 and 83 (Fig. F229, F230). Mineralogically, it appears to be an oxide-rich (~20 vol%) tonalite, but its chemical composition is that of an evolved FeTi basalt (49% SiO2, 4% MgO, 18% FeO, and 4% TiO2) and the rock name “diorite” is more appropriate. Primary minerals include abundant quartz (20%–25%), abundant (>20%) interstitial Fe-Ti oxides, a primary mafic phase that was probably hornblende, and plagioclase, both as zoned subhedral tabular prisms and as granophyric intergrowths with quartz (Fig. F231A). Primary features are obscured by pervasive alteration of mafic phase(s) to actinolite and near-complete replacement of plagioclase (Fig. F231).

Units 1256D-83 through 87: medium-grained disseminated oxide gabbros

Units 1256D-83 through 87 are lithologically very similar, distinguished from one another by relatively minor textural differences. All are predominantly medium grained with primary magmatic features mostly obscured by pervasive alteration and only patchy primary domains remaining (Fig. F232). Textures in network domains are typically subhedral, inequigranular, and seriate with ophitic or subophitic patches defining the dark-colored domains (Figs. F210, F233). The size of clinopyroxene oikocrysts and the continuity of network domains are key variables that help to define units.

Unit 1256D-84 is essentially identical to Unit 1256D-83 and is defined separately only because the two units are not contiguous in the core.

Unit 1256D-85 is distinguished by the development of a distinct patchy texture (Fig. F205). The dark domains include clinopyroxene oikocrysts 1 cm or more in diameter, and strands of the leucocratic network domains are 1–5 mm wide.

Unit 1256D-86a is defined on the basis of a gradual textural change from patchy in Unit 1256D-85 to more equigranular with readily visible ophitic clinopyroxenes in Unit 1256D-86.

Unit 1256D-86b is a medium-grained oxide gabbro that is coarser and more strongly altered than Unit 1256D-86a (Fig. F234). It appears to have intruded into Unit 1256D-86a.

Unit 1256D-87 contains much larger clinopyroxene domains (as large as 20 mm) than those in Unit 1256D-86. Each domain is composed of several anhedral oikocrysts. Coronas of hornblende that include vermicular (symplectite-like) reaction textures around clinopyroxene may be magmatic in origin. Similar coronas are present in Unit 1256D-85.

Unit 1256D-88: medium-grained disseminated oxide gabbro

Unit 1256D-88 is defined by the first appearance downhole of diffuse centimeter-scale coarse-grained patches that are highlighted by pale (altered) plagioclase, higher oxide mineral contents, and the absence of poikilitic clinopyroxene. The patches are scattered within medium-grained gabbro that continues from Unit 1256D-87. The coarse patches appear igneous in origin, but they appear more strongly altered than the background gabbro, perhaps because of the paler plagioclase. Unit 1256D-88 is also distinguished by the appearance of scattered olivine as small, highly altered interstitial grains with dark, oxide-rich alteration halos (Fig. F235).

Unit 1256D-89: medium coarse–grained olivine- and orthopyroxene-bearing oxide gabbro

Unit 1256D-89 is distinguished from Unit 1256D-88 by its smaller grain size, high but variable olivine contents of up to ~20%, and the first appearance of pargasitic amphibole in the high-grade metamorphic domains. In other respects, Unit 1256D-89 gabbros resemble those of the other Gabbro 1 units.

Unit 1256D-89 is cut by a pair of narrow coarse-grained oxide gabbro dikes (Unit 1256D-89b), mineralogically similar to the diffuse network domains of Units 1256D-83 through 87 (Fig. F236).

Unit 1256D-90: dike screen: fine-grained metabasalt

Unit 1256D-90 consists of fine-grained (meta)basalts very similar to those at the base of the sheeted dike interval. The igneous texture appears to have been overprinted by a granular metamorphic texture cut by several thin gabbroic and leucocratic dikes. At its lower boundary, Unit 1256D-90 is intruded by gabbronorite from Gabbro 2 (Fig. F237). Approaching the contact, secondary granular orthopyroxene appears to progressively replace clinopyroxene (Fig. F238), suggesting that orthopyroxene formed by prograde reactions that were more intense in the boundary zone.

Units 1256D-90b through 90f are small narrow dikes of medium-grained quartz gabbro (Fig. F239) (Units 1256D-90b, 90d, and 90e) and fine-grained tonalite (Units 1256D-90c and 90f).

Gabbro 2 (Units 1256D-91 through 95)
Units 1256D-91 through 93: medium-grained gabbronorite and orthopyroxene-bearing gabbro

Gabbro 2 appears to be a single intrusion. Units 1256D-91 and 93 are its upper and lower marginal zones, characterized by xenolithic inclusions of fine-grained basaltic material apparently stoped from the adjacent (meta)basaltic dike screens. Units 1256D-91 and 93 are gabbronorites, reflecting higher orthopyroxene abundances near the gabbro margins. Unit 1256D-92 is an orthopyroxene-bearing gabbro. Throughout Gabbro 2 and in the adjacent basaltic units, orthopyroxene occurs in several forms (Fig. F240), both magmatic and metamorphic, emphasizing that the interval including and beneath the dike screen is an important interface region where the distinction between magmatic and metamorphic processes is difficult to discern.

Unit 1256D-91 intrudes Unit 1256D-90, the upper dike screen, along a complex subvertical contact (interval 312-1256D-230R-1, 15–24 cm) (Fig. F237). Along this contact, medium-grained gabbro invades the metabasalt screen, surrounding and detaching fragments of metabasalt. The metabasalt fragments appear slightly pink in hand specimen, reflecting the presence of orthopyroxene in the groundmass. The predominant primary minerals are plagioclase, clinopyroxene, orthopyroxene, and oxide forming fine-grained inequigranular to poikilitic textures. High-grade granoblastic metamorphic domains rich in orthopyroxene but otherwise similar to those in the deeper basaltic units also occur.

Unit 1256D-92 (Fig. F241) is defined by the absence of fine-grained basaltic xenoliths. The medium- to fine-grained orthopyroxene-bearing gabbros are generally similar to those of Unit 1256D-91a, with heterogeneous subhedral inequigranular seriate to poikilitic textures. Orthopyroxene is less abundant than in the marginal units. Olivine, which is rare or absent in Units 1256D-91 and 93, is present in small amounts but almost completely altered.

Unit 1256D-93, at the lower margin of Gabbro 2, is defined by the presence of rounded basalt xenoliths apparently stoped from the underlying Unit 1256D-94, but the intrusive contact was recovered in only one small piece. It is very heterogeneous, and orthopyroxene continues to occur in a variety of magmatic and metamorphic forms (Figs. F242, F243, F244, F245).

Unit 1256D-94: fine-grained gabbronorite or metabasalt

Unit 1256D-94 is orthopyroxene-rich, recrystallized, cryptocrystalline to fine-grained rock of basaltic composition. It is extensively recrystallized with well-developed granoblastic textures, but in some areas, especially away from its contact with Unit 1256D-93, it retains an igneous intergranular texture (Figs. F246, F247B). There is some ambiguity as to the most appropriate rock name, either metabasalt or fine-grained gabbronorite. The latter term has been applied to similar lithologies from the root zone of the sheeted dike complex in the Oman ophiolite (Nicolas and Boudier, 1991; Boudier et al., 2000).

The primary minerals, plagioclase, clinopyroxene, orthopyroxene, and Fe-Ti oxide, can be inferred from regions of relict fine-grained intergranular texture. Occasional orthopyroxene-rich poikilitic or poikiloblastic domains (Fig. F247A) distinguish this unit from other Expedition 312 metabasaltic rocks. Elsewhere, isolated orthopyroxene prisms within the groundmass can be interpreted to be of magmatic origin. As for Gabbro 2, orthopyroxene appears to have formed by a mix of igneous and metamorphic processes.

In the lower part of the unit, away from the unit boundary, large orthopyroxenes are absent and plagioclase is less dusty in appearance (Fig. F246B), suggesting that metamorphism was less intense away from the gabbro contact. Smaller prismatic orthopyroxene is still present, however, and it cannot be determined with certainty if its origin is igneous or metamorphic.

Unit 1256D-95: crosscutting basalt dike

Unit 1256D-95 is a cryptocrystalline aphyric basalt distinguished from Unit 1256D-94 by its lighter gray color and finer grain size. Its most distinctive features are the presence of abundant pink clinopyroxene, possibly titanaugite, and its relatively low primary oxide content (Fig. F248). It is also distinct from the other units of both dike screens in its relatively low metamorphic grade and degree of alteration (texture Type 3–4). This crosscutting, relatively enriched basalt dike or sill must have been intruded significantly later than the gabbroic units.

Geochemistry

A total of 46 whole rock samples from Hole 1256D were analyzed for major and trace element concentrations by ICP-AES during Expedition 312 (see the “Methods” chapter). Whenever possible, at least one representative fresh rock was selected from each igneous unit and analyzed in order to obtain the unit’s magmatic composition at the time of emplacement. These rocks have been classified as basaltic dikes (25), gabbros (15), quartz-rich oxide-diorite (1), or trondhjemite (1) in addition to basalts (3) and an unaltered dolerite from a ghost core (see “Igneous petrology”). Expedition 312 dikes are variably altered, as indicated by LOI values up to 2 wt%. Four samples with visibly extensive alteration and LOI >2 wt% were classified as altered and were excluded from general data description and petrogenetic interpretation. These samples are considered below in “Geochemistry of altered samples.” Rocks recovered from shallower depths in the hole, particularly those from Expedition 309, were described as having positive correlations between grain size and alteration extent (see “Geochemistry” in “Expedition 309”). This trend does not continue deeper in the hole, and, if anything, the opposite is commonly true in Expedition 312 rocks. Therefore, during Expedition 312, representative samples were chosen from each igneous unit that were the most free from visible alteration (veins, halos, and patches) without regard to their grain size. When coarser grained rocks were sampled, a larger volume was taken (up to 30 cm3 instead of 10 cm3) to ensure the analysis was representative of the whole rock.

All ICP-AES major element analyses are anhydrous and have been normalized to 100 wt%. All analyses obtained during Expedition 312 had prenormalized major element totals between 98.4 and 101.1 wt%, indicating major element analyses of high quality. Iron is reported as FeOT and calculated to be 0.8998 of Fe2O3. Major and trace element compositions and LOI data are presented in Table T37, along with ratios of selected elements. During the shipboard ICP-AES runs, reproducibility and accuracy were monitored by multiple analyses of shipboard standards (BAS-140, BAS-148, and BAS-206) from previous ODP drilling legs and a sample from the first ghost core recovered during Expedition 312, BAS-312. The relative standard deviation estimated from multiple analyses of these standards is ±2% for major elements except for K2O and P2O5 because of low sample concentrations close to background levels. Trace element analyses have reproducibility of ±5% except for Sr and Ba, which have reproducibilities of ±15%. Estimated errors on averaged standard measurements can be found in Table T14 in the “Methods” chapter. Details of the analytical procedure, instrument running conditions, and sampling protocol can be found in “Geochemistry” in the “Methods” chapter.

Geochemistry of Expedition 312 basaltic dikes, gabbros, and leucocratic rocks

Downhole geochemical variations are presented along with their igneous unit and rock type in Figures F249 and F250. A thorough overview of Leg 206 and Expedition 309 geochemical data from shallower depths in Hole 1256D is presented in “Geochemistry” in “Expedition 309.”

Dikes

Geochemical compositions of Expedition 312 dikes are variable and do not define obvious trends downhole. This contrasts with the downhole observations of rock geochemistry from shallower depths in the hole. Expedition 312 dikes lie within the range of major element compositions found in the overlying rocks; for example, Expedition 312 dikes have SiO2 = 49.7–52.1 wt%, TiO2 = 0.97–2.3 wt%, and MgO = 5.3–8.5 wt% compared to 48.0–55.5, 1.0–3.1, and 4.9–9.7 wt%, respectively, for Leg 206 (Wilson, Teagle, Acton, et al., 2003) and Expedition 309 rocks. This is also true for most trace element concentrations (Sc, Co, Zr, Y, Sr, and Ba), but new minima for Zn and V concentrations and maxima for Ni and Cr abundances in Hole 1256D samples are found in Samples 312-1256D-189R-1 (Piece 10, 66–68 cm) and 190R-1 (Piece 2, 14–17 cm). These dikes are from igneous Unit 1256D-77 and are associated with comparatively primitive Mg#s of ~61. Although these dikes are not the least evolved of all rocks recovered from Hole 1256D, they are the least evolved of the recovered dikes (in terms of Mg#), have the highest concentrations of compatible trace elements (Ni and Cr), and are the least enriched in some incompatible trace elements (Zn and V). Overall, they are important samples, as they have the most primitive magma compositions of the Site 1256 sheeted dike complex. The most fractionated dike analyzed is Sample 312-1256D-176R-2 (Piece 4B, 22–25 cm) from igneous Unit 1256D-72. This dike has a low Mg# (38) and high concentrations of Zr (118 ppm), Y (37 ppm), and V (422 ppm) but is not particularly distinctive in other element compositions. The less fractionated dike samples from Unit 1256D-77 correspond to a thick igneous unit, whereas the more evolved dike of Unit 1256D-72 is from a thin igneous unit. Fractionation downhole relative to interpreted dike thickness is addressed in “Igneous petrology” and Figure F201.

A leucocratic dike crosscuts the basalt sheeted dikes in Unit 1256D-80b (Sample 312-1256D-212R-1 [Piece 7, 29–32 cm]). This dike is quartz rich with high SiO2 (72 wt%), Zr (840 ppm), and Y (50 ppm) and low MgO (1.1 wt%), CaO (4.9 wt%), and Ni (9 ppm). This dike falls on the extreme fractionated end of most geochemical trends for Site 1256 rocks. Petrographically and chemically, this rock is a trondhjemite, having high SiO2 and Na2O but low K2O, TiO2, and FeO.

Gabbros

The plutonic section recovered two gabbro sequences (Gabbros 1 and 2) intruded into dike screens (Fig. F202). These gabbros have highly variable bulk geochemical compositions. The uppermost gabbros (Gabbro 1; Units 84–87; ~1412–1422 mbsf) have geochemical characteristics similar to the overlying dikes. These gabbros are fractionated and have MORB chemistry, with MgO ranging from ~7 to ~8 wt% and Zr ranging from ~47 to ~65 ppm. Conversely, deeper in Gabbro 1 (Units 88–89; ~1436–1451 mbsf), gabbros are significantly less fractionated with high MgO (11.4 wt%) and Ni (~200 ppm) and low TiO2 (0.75 wt%) and Zr (~40 ppm) present in Sample 312-1256D-223R-2, 41–48 cm. There are general downhole trends in Gabbro 1 of increasing MgO, CaO, Ni, and Cu and decreasing FeO, Zr, and Y. These trends correspond to decreases in modal abundances of oxides in the gabbros. Iron contents in the upper gabbros of Gabbro 1 are low relative to their MgO contents.

A quartz-rich oxide diorite was recovered high in Gabbro 1 (1411.32 mbsf; interval 312-1256D-214R-1, 42–47 cm). The geochemistry of this unit suggests it is a highly fractionated equivalent of Expedition 312 dikes and gabbros. It is similar in bulk composition to an evolved FeTi basalt, containing high concentrations of Fe and Ti (FeO = 17.44 wt% and TiO2 = 4.12 wt%). Although it is quartz bearing, it has relatively low silica contents (SiO2 = 49.6 wt%), a result of its high oxide abundance.

The uppermost rocks of Gabbro 2 are fractionated with MgO contents of 6.1 wt%, but lower in the sequence MgO reaches 9.3 wt%. Trends present in downhole plots also suggest the extent of fractionation decreases with depth in Gabbro 2. For example, TiO2 is 2.5 wt% at the top and 1.2 wt% at the base and FeO is 16.4 wt% at the top and ~10 wt% at the bottom of Gabbro 2. Compositions of Gabbro 2 vary from the fractionated end of the sheeted dike complex to values more primitive than those found in any of the dikes.

Petrogenetic interpretation of geochemical trends in Expedition 312 rocks

The dikes and uppermost gabbros recovered during Expedition 312 are fractionated, MORB-type rocks spanning the basalt to basaltic andesite fields on a plot of total alkalis versus silica (Fig. F251). Compositional variation in dikes sampled during Expedition 312 is large and covers the entire range present in the rocks at shallower depths in Hole 1256D. Dikes sampled during Expedition 312 demonstrate clear trends in plots of major and trace elements versus MgO wt% and span ranges similar to Expedition 309 dikes (Figs. F252, F253). End-member compositions to these trends occur in the lowermost gabbros from Gabbro 1 (primitive end; Samples 312-1256D-223R-1, 35–42 cm, and 223R-2, 41–48 cm) and the trondhjemite dike and the quartz-rich oxide diorite (fractionated end; Samples 212R-1, 29–32 cm, and 214R-1, 42–47 cm).

The dominant trends present in Hole 1256D rocks can be explained by fractional crystallization (Fig. F254). The fractionating assemblage in Hole 1256D rocks is clinopyroxene and plagioclase, as is expected for relatively evolved basaltic magmas. This is well displayed in a plot of Sc/Y versus Sr/Y (Fig. F255). Fractionation of plagioclase preferentially decreases Sr/Y because Sr is more compatible in the plagioclase crystal lattice than Y; removal of clinopyroxene lowers Sc/Y, as Sc is more compatible in clinopyroxene than Y. The least fractionated dikes have high Sc/Y (~1.6) and moderate Sr/Y (~5.2). The least fractionated gabbros have the highest Sc/Y (~2.1) and Sr/Y (4–4.6). The remaining dikes follow a clear trend toward low Sc/Y (~0.8) and Sr/Y (~1.8), indicating that clinopyroxene and plagioclase fractionation is the principal influence on magma chemistry. Additional evidence for fractional crystallization is detailed below in the PCA results.

The overall low FeO found in many Gabbro 1 rocks, the decrease in oxide and increase in olivine modal abundances, and the general geochemical trends toward higher MgO and Ni and lower FeO and Zr suggest that the lowermost gabbros of Gabbro 1 may be cumulate rocks. Therefore, the unusually high MgO of ~11.4 wt% in one of these gabbros (Sample 312-1256D-223R-2, 41–48 cm) should be interpreted with caution. Examination of the thin section from this rock provides evidence for strong magmatic disequilibrium, including zones of olivine adjacent to hornblende, deeply embayed plagioclase phenocrysts, and quartz. It is possible that this rock was originally a basalt with a large amount of accumulated olivine that was strongly altered to produce the disequilibrium mineral phases observed (see “Igneous petrology”). Alternatively, this gabbro may be a residue produced when melt was extracted during formation of the sheeted dike complex. Regardless, based on the above observations, it is reasonable to conclude that the high MgO value is a result of crystal accumulation rather than a primary melt signature.

Vari-textured gabbros are observed in frozen melt lenses located between the sheeted dike complex above and the cumulate crystal mush pile below in many ophiolites. For example, the isotropic microgabbros located beneath the dike root zone of the Oman ophiolite are interpreted to be crystallized liquids (MacLeod and Yaouancq, 2000). By analogy, the granoblastic dikes that form the base of the sheeted dike complex in Hole 1256D may represent a dike root zone similar to that described in the Oman ophiolite. It follows that below the sheeted dike complex, Hole 1256D gabbros may represent part of the fossilized melt lens. The trends in Gabbros 1 and 2 are clearly toward more mafic compositions downhole, indicating in situ fractional crystallization.

Principal component analysis

PCA is a method for examining correlated geochemical variation within large sets of data. This method allows isolation of linear geochemical trends that correspond to data variation both by reinforcing the correlated signals from different elements and by improving the signal-to-noise ratio of the trends that have been isolated. Improvement of the signal-to-noise ratio is particularly desirable when dealing with potentially noisy shipboard ICP-AES data. Details of PCA have been described by numerous authors, most recently by Albarède (1995) and Maclennan et al. (2001). The approach of Albarède (1995) was used in this report: the matrix of linear correlation coefficients for all analyzed elements was calculated first, followed by calculations of the principal components from the eigenvalues and eigenvectors of the correlation matrix (Table T38).

The primary igneous signal was isolated from alteration effects by calculating the principal components for samples with <1 wt% LOI. Poor correlations were found between Si, K, and Ba and other elements, so these elements were then excluded from analysis to prevent analytical outliers from dominating the PCA. The first principal component likely reflects variable extents of fractional crystallization of gabbro and is consistent with the above interpretation of the data. More negative values for the first principal component correspond to increasing degrees of magmatic evolution, whereas positive values indicate lower extents of fractionation. There is a clear large-scale decrease in the degree of fractionation from the uppermost lavas (~250 mbsf) toward the transition zone (~1000 mbsf) (Fig. F256). Large variations are present between the transition zone and the sheeted dike complex over small depth intervals, with no obvious trend with depth. Superimposed on these first-order variations, it is possible to discern coherent evolving sequences (e.g., from 430 to 280 mbsf or 700 to 600 mbsf) and bimodality within some intervals, such as between 430 and 580 mbsf. It appears from the similar spread of first principal component values in dikes and lavas that the dikes were likely the feeders for most of the lavas.

Evidence for mantle heterogeneity

Ratio-ratio plots of highly incompatible elements minimize the effects of variation in extent of fractionation and/or extent of partial melting and permit examination of mantle source characteristics. Fractional crystallization of clinopyroxene and plagioclase from a basaltic system should result in a positive slope in these diagrams. Trends present in ratio-ratio plots of Hole 1256D samples cannot be entirely explained by crystal fractionation alone and intimate a heterogeneous source (Fig. F257). For example, in Zr/Y versus Ti/Y or V/Y, variable values for Zr/Y at a given Ti/Y or V/Y exist, suggesting a difference in their mantle source. Source heterogeneity is also the favored explanation for the presence of varying Zr/Y for a given MgO content, as the latter decreases by definition during the crystallization of clinopyroxene but Zr/Y should correspondingly increase (Fig. F254). Finally, a heterogeneous mantle source for Hole 1256D rocks is further supported in a plot of Sc/Y versus Sr/Y, as some rocks fall well outside of the main fractional crystallization trend (Fig. F255). There is a significant amount of work supporting the idea that the upper mantle, although depleted on average, is heterogeneous on a small scale (e.g., Sinton et al., 1991). Evidence for a heterogeneous mantle source from Hole 1256D is not unexpected, even at a superfast spreading center where magma chambers may be steady state and well mixed on a large scale (e.g., Batiza and Niu, 1992).

Geochemical comparison between Site 1256, modern East Pacific Rise, and Site 504

The dikes, massive basalts, sheet flows, and ponded lavas from Site 1256 are less depleted than other MORB rocks, such as those recovered during drilling at Site 504 in 6.9 Ma rocks formed at the Costa Rica Rift (Figs. F254, F257) (Autio et al., 1983; Natland et al., 1983; Emmermann, 1985; Tual et al., 1985). Superfast spreading along the northern EPR (220 mm/y) (Wilson, 1996) has produced less incompatible element–depleted primary magmas relative to magmas formed at the intermediate spreading Costa Rica Rift (45 mm/y at 98°W) (DeMets et al., 1990). The possible causes for this include a more depleted mantle source or larger degrees of partial melting in the latter. Hole 1256D rocks have geochemical characteristics comparable to, but with somewhat lower incompatible element abundances than, modern MORB formed along the northern EPR between 5° and 10°N (Fig. F76) (Langmuir et al., 1986; C.H. Langmuir, unpubl. data, www.petdb.org, 1999), close to the extrapolated location of formation of Site 1256 ~15 m.y. ago. Hole 1256D rocks generally have higher Ti/Zr and lower Zr/Y, Ti/Y, and V/Y (Figs. F254, F257) relative to modern EPR basalts.

Two main explanations exist for the differences in the incompatible trace element ratios of the modern EPR and Hole 1256D basalts: (1) changes in the melt supply or in the stability of the magma chamber (e.g., Regelous et al., 1999) or (2) off-axis magmatism. A decrease in the melt supply from the mantle would tend to produce lower MgO contents and more enriched magmas. Variations in both the source of magmas and the supply of those magmas that fed the melt lens can explain much of the geochemical range observed in Site 1256 rocks. The downhole trends observed at shallower depths in Hole 1256D (e.g., 430–280 and 700–600 mbsf) and in the Gabbro 1 sequence appear to be related to distinct cycles of fractionation and magma chamber recharge. It has been proposed on the basis of both geochemistry (Reynolds et al., 1992) and magnetic data (Carbotte and Macdonald, 1992) that axial magma chambers are recharged cyclically, but without age data on the specific lithologic zones in Hole 1256D this cannot be confirmed.

Alternatively, off-axis magmatism may be recorded in Hole 1256D rocks. Evidence for probable off-axis magmatism is found in the ponded lavas from the top of the hole (see “Igneous petrology” in “Expedition 309”). The ponding of these lavas requires significant topography, where magmas can stagnate and crystallize, and this is developed at the modern EPR 5–10 km off axis. These lavas may have formed from magmas that were erupted at the axis and flowed onto the ridge flanks or, alternatively, they could have formed from an off-axis vent. Off-axis volcanism, sampled in Hole 1256D, may have played a role in the chemical differences between this site and EPR rocks. Off-axis volcanism can tap magmas from deeper in the melting column (e.g., Batiza and Niu, 1992), in which case, smaller extents of partial melting are expected, consistent with the observation that lava pond rocks in Hole 1256D are generally more evolved than the axial lavas of the sheet and massive flows.

Geochemistry of altered samples

Analyses of highly altered samples, and those having >2 wt% LOI, are presented in Table T39. Three altered dikes from Expedition 312 were selected for geochemical analysis (Samples 312-1256D-184R-1, 98–104 cm, 187R-1, 93–98 cm, and 192R-1, 11–13 cm). Bulk geochemistry of these samples can be compared to similarly evolved, relatively fresh samples to obtain a general indication of elemental mobility. This method of determining element mobility is more effective for homogeneous glassy samples than in coarser grained rocks because coarser grained rocks are not uniformly altered and may contain halos, veins, or other local and/or focused alteration patches. The sample selected as representative of an unaltered, similarly fractionated equivalent (Sample 309-1256D-110R-1, 58–66 cm) is from a massive basalt flow drilled during Expedition 309, and its elemental composition may be found in “Geochemistry” in “Expedition 309.” Elemental data for altered samples are normalized to elemental data for the least altered sample to place some constraints on element mobility (Fig. F258). Elements that are enriched (>1) or depleted (<1) in this diagram may be so because of the addition of hydrothermally delivered minerals or removed during hydrothermal seawater circulation. The immobile elements Ti and Zr exhibit no significant change, showing that the fresh rock selected for normalization is reasonable. Mobile elements in the altered dike samples include P, Na, Zn, Cr, and Cu. Na is enriched in all three samples. Preferential depletion in Zn, Cr, and Cu (below detection limits in two of the altered samples) also occurs in all altered samples. The discrepancy in Cr contents in the altered and fresh rocks likely reflects a difference in rock composition; however, Zn and Cu loss probably indicates leaching by hydrothermal fluids at high temperature (>350°C). Dike Samples 312-1256D-184R-1, 98–104 cm, and 187R-1, 93–98 cm, have elevated K and Ba values relative to the least altered sample, whereas dike Sample 192R-1, 11–13 cm, is depleted in these elements. This apparent discrepancy may be caused by variable alteration extent of magmatic plagioclase in the dikes because plagioclase can incorporate trace amounts of K and Ba in its crystal lattice. The slight loss of Ca is consistent with the albitization of plagioclase and gain of Na. Although Mg is taken up by basalts during low-temperature exchange with seawater, it does not appear to be so at depths corresponding to the sheeted dike complex (>1060 mbsf), as demonstrated by normalized values of ~1 in the altered rocks. In summary, alteration of dikes in the sheeted dike complex is likely the result of the presence of evolved (Mg depleted) seawater fluids at these depths.

Alteration

Hydrothermal alteration by seawater-derived fluids of dikes and gabbros intruded at mid-ocean ridges is a major mechanism of heat and chemical transfer from the oceanic basement to the oceans. The location, geometry, and intensity of hydrothermal circulation profoundly influences styles of magmatic accretion as well as chemical fluxes and mineralization of the ocean crust. Magmatic, hydrothermal, and tectonic processes at the ridge axis are closely interrelated, and the description and interpretation of these phenomena require comprehensive integration of these observations.

Observations during Expedition 312 form a continuum with core descriptions made during Expedition 309 (see “Ocean crust formed at a superfast spreading rate: deep drilling of ocean basement in Hole 1256D”) and Leg 206 (Wilson, Teagle, Acton, et al., 2003), and the dikes will be considered within the context of the complete sheeted dike complex penetrated by Hole 1256D. The extent and distribution of alteration were recorded section by section on the VCDs (see Fig. F2 in the “Methods” chapter) for each igneous unit. Bulk rock alteration observable in hand specimen is described quantitatively in the alteration logs (see 312ALT.XLS and PLUTLOG.XLS in “Supplementary material”). The composition and distribution of veins, breccias, and associated alteration halos are recorded in the vein log (see 312VEIN.XLS in “Supplementary material”). Macroscopic observations were supplemented and calibrated through detailed thin section investigations (Table T40; see “Core descriptions”) and XRD analyses of vein fillings and breccia cements (Table T41). Vein and mineral occurrences and the distribution of alteration styles within the sheeted dikes and plutonic section of Hole 1256D are summarized in Figures F259, F260, F261, F262, and F263.

Sheeted dikes

The sheeted dikes from an intact section of ocean crust (interval 309-1256D-129R-1, 0 cm, through 312-1256D-213R-1, 52 cm; 1060.9–1406.62 mbsf) were penetrated for only the second time in the history of scientific ocean drilling during Expedition 309/312, providing the first comparison with dikes recovered from Hole 504B (see Alt et al., 1996a). Here, we briefly summarize observations of hydrothermal alteration of the sheeted dikes in the cores that directly overlie the interval drilled during Expedition 312.

Expedition 309 recovery

Approximately 195 m of massive basalts with common subvertical intrusive contacts was drilled during Expedition 309, and these rocks are interpreted to be the upper portion of the sheeted dike complex at Site 1256 (Sections 309-1256D-129R-1 through 170R-3; 1060.9–1255.1 mbsf). The rocks are slightly to moderately altered, dark gray to dark green-gray basalts in which groundmass igneous phases are partially altered to greenschist-facies secondary minerals. Groundmass and phenocryst plagioclase are partially altered to chlorite and/or albite. Clinopyroxene is partly replaced by chlorite, although actinolite is also common from 1118.8 to 1255.1 mbsf. Throughout the Expedition 309 section, groundmass clinopyroxene in the sheeted dikes is commonly dusty, corroded, and altered to microscopic aggregates of intergrown actinolite needles and secondary magnetite. Transmitted electron microscope observations of similar dusty clinopyroxenes from Hole 504B suggest that intermediate “pyribole” crystal structures are present (C. Laverne, unpubl. data).

Hydrothermal veins are common throughout the Expedition 309 sheeted dikes (~25 veins/m) (Fig. F259; Table T42), with chlorite the most abundant vein component. Quartz, pyrite, chalcopyrite, actinolite, prehnite, laumontite, and calcite are also common vein components. Crosscutting relationships indicate that groundmass replacement and vein filling by chlorite, titanite, albite, actinolite, and pyrite are relatively early, and this pervasive alteration can be overprinted by hydrothermal veins composed of quartz, chlorite, epidote, pyrite, chalcopyrite, and rare sphalerite. Late-stage crosscutting assemblages that probably formed at lower temperatures (100°–250°C) include anhydrite, prehnite, laumontite, and calcite. Many veins have well-developed alteration halos in which mixtures of chlorite, albite, actinolite, titanite, quartz, pyrite, calcite, and prehnite replace plagioclase and clinopyroxene and fill interstitial pore space.

Hydrothermal alteration is most spectacularly manifest by two phenomena: (1) centimeter-scale hydrothermal alteration patches and (2) mineralized dike margins. Alteration patches comprise centimeter-scale zones of 100% hydrothermal minerals, most commonly quartz, prehnite, laumontite, chlorite, anhydrite, and calcite, either replacing basalt or filling pore space, surrounded by dark chloritic halos in which there is almost complete replacement of groundmass phases (e.g., Fig. F129). Many subvertical dike margins encountered are disrupted by a complex vein network that brecciates the chilled contacts, with intense hydrothermal recrystallization of the surrounding rock to chlorite, actinolite, quartz, pyrite, and chalcopyrite (e.g., Fig. F61). Crosscutting anhydrite veins are common.

Expedition 312 recovery

Rocks recovered from Hole 1256D during Expedition 312 (interval 312-1256D-172R-1, 0 cm, through 213R-1, 52 cm; 1255.1–1406.62 mbsf) are moderately to completely altered. Alteration occurs by replacement of igneous groundmass phases and phenocrysts and is most intensely developed in alteration patches, breccias, veins, vein networks, and alteration halos.

Background alteration

The basalts of the sheeted dikes cored during Expedition 312 are dark gray to dark green-gray in hand specimen and moderately altered (Fig. F264). Disseminated pyrite is a common accessory phase. Fine- to medium-grained basalts commonly appear greener than microcrystalline rocks (Fig. F264C). Background alteration most commonly occurs through the pseudomorphic replacement of igneous phases by secondary minerals, and igneous textures are generally preserved. Small patches of groundmass recrystallized to microcrystalline granular secondary pyroxene and plagioclase are observed in basalts from Core 312-1256D-190R and below (1343 mbsf).

In line with observations from Expedition 309, groundmass clinopyroxene is moderately to highly altered, appearing dusty and corroded, with wispy actinolite and specks of secondary magnetite (Fig. F265). Pyroxene replacement takes place in situ, with basalts retaining their igneous textures. In zones where alteration is more developed, clinopyroxene is replaced by actinolite with minor chlorite. Chlorite is the dominant mafic alteration phase to ~1310 mbsf (Section 312-1256D-182R-1), below which actinolite becomes more abundant, in association with secondary magnetite. Below 1348 mbsf (Section 312-1256D-192R-1), clinopyroxene is commonly replaced by a core of brown-green pleochroic hornblende surrounded by actinolite. The alteration of groundmass plagioclase is generally less advanced than that of the coexisting clinopyroxene, and plagioclase can be replaced by either albite or by chlorite/​actinolite. Recrystallization occurs along cleavage planes and around crystal margins. Glass inclusions within plagioclase crystals are mostly replaced by chlorite or actinolite. Secondary calcic plagioclase, occurring as inclusion-rich rims of igneous plagioclase, is present in brown-green hornblende-bearing rocks below ~1348 mbsf.

Most basalts from Hole 1256D sheeted dikes are aphyric. Where present, rare phenocrysts are most commonly glomerocrystic plagioclase with lesser amounts of clinopyroxene. Phenocrysts are altered in a similar style to the host groundmass. Plagioclase phenocrysts are partially recrystallized to albite, chlorite, and quartz. Late-stage prehnite and laumontite are also present. Clinopyroxene phenocrysts are recrystallized to dusty clinopyroxene/​actinolite and magnetite in the upper dikes but more extensively replaced by actinolite, locally with minor chlorite or hornblende, deeper in the dikes (Fig. F266). Olivine phenocrysts have not been observed in the sheeted dikes, but rare phenocrysts with outlines reminiscent of relict olivine are completely replaced by chlorite, talc, and magnetite.

Veins and vein halos

Veins with or without alteration halos are abundant features throughout the Expedition 312 sheeted dikes (Figs. F259, F260, F264, F267), with an average of ~35 veins/m (Table T42). There is a trend of increasing veins per meter from ~1325 to 1411 mbsf, but this most likely reflects low core recovery in this interval. Although there is not a strong relationship between core recovery and observations of the number of veins per meter, the highest frequency of veins generally occurs when core recovery is very low (Fig. F268). Low core recovery in the lower parts of the sheeted dikes is probably responsible for the apparent increase in the number of veins per meter with depth (Fig. F259). Veins are typically 0.1–1.5 mm wide and rarely 2–5 mm wide. Alteration halos 1–5 mm wide are commonly associated with veins. Chlorite is the most abundant vein component throughout the Hole 1256D sheeted dikes, although actinolite veins become increasingly abundant below ~1330 mbsf and are dominant in the lowermost dikes. Sulfides (principally pyrite) are common, albeit irregularly distributed, vein constituents throughout the dike section. Quartz, most commonly occurring in ~1 mm wide veins with chlorite and pyrite, is present throughout the dike section but most abundant from ~1250 to 1350 mbsf. Anhydrite veins, common in the upper dikes (above 1200 mbsf), are rare in the lower reaches of Hole 1256D (Fig. F259).

There are many different arrangements of vein filling, and complex crosscutting relationships are exhibited. Chlorite is present in most veins from Cores 312-1256D-172R through 179R (1255–1295 mbsf), commonly occurring with quartz, pyrite, and titanite and more rarely with epidote, prehnite, laumontite, secondary magnetite, actinolite, or anhydrite. Rare veins also include chalcopyrite or sphalerite. Chlorite, quartz, actinolite, titanite, pyrite, chalcopyrite, and epidote generally appear to have precipitated during the same phase of hydrothermal mineralization, but quartz is commonly replaced by later laumontite and/or prehnite, and all these phases can be crosscut by anhydrite, laumontite, or calcite veins. Actinolite is the most abundant vein filling from Core 312-1256D-179R (~1290 mbsf) to the bottom of the dikes, although subordinate chlorite is commonly present. Actinolite occurs in veins with chlorite, quartz, prehnite, albite, magnetite, titanite, and rare epidote. Blue-green amphibole is observed in Sample 312-1256D-190R-1, 14–16 cm, and pleochroic brown-green hornblende is present with actinolite in veins in interval 192R-1, 11–13 cm (~1348 mbsf).

Many veins have alteration halos that are clearly visible in hand specimen, with a 1–2 mm wide dark green border commonly flanked by a wider (2–5 mm) light gray outer halo (Figs. F264, F267). Igneous minerals are more highly recrystallized in these halos compared to the background basalts, with titanomagnetite commonly the only remaining igneous phase. Plagioclase is partially to completely replaced by albite, chlorite, quartz, and actinolite. Clinopyroxene is completely replaced by dusty actinolite-magnetite intergrowths or by actinolite and minor chlorite. Below ~1346 mbsf, alteration halos are different. Clinopyroxene may be replaced by brown-green hornblende associated with green actinolite. Pyrite is commonly disseminated in the halos, igneous titanomagnetite is replaced by titanite, and secondary magnetite is common. Instead of the dusty titanite observed in the upper dikes (to ~1290 mbsf), euhedral or subhedral grains of titanite are present in Cores 1256D-312-179R through 190R.

Alteration patches

Centimeter-scale isolated zones composed entirely of greenschist-facies hydrothermal minerals are common features in the sheeted dikes to ~1350 mbsf. These patches are commonly elliptical (see “Structural geology”), appear unrelated to hydrothermal veins, and comprise central cores surrounded by centimeter-scale halos of highly to completely altered wallrock. Secondary minerals in the central cores may have grown into open space or have completely obliterated preexisting igneous minerals or textures. Causes for the focusing of fluid–rock reactions and hydrothermal precipitation into these alteration patches remain unclear but may be initiated around recrystallized phenocrysts or develop in small areas of elevated primary porosity or mesostasis (see “Structural geology”). The processes of patch initiation are recorded in interval 312-1256D-176R-2, 22–25 cm, where 5 mm patches have developed in the slightly coarser grained regions of the basalt in which there is albitization of plagioclase and partial replacement by chlorite. Clinopyroxene is recrystallized to dusty clinopyroxene/​actinolite with common protruding actinolite needles (Fig. F269).

Within larger, centimeter-scale alteration patches, the central core of the patch is most commonly filled by quartz, chlorite, and pyrite grown into open space (Fig. F269), although these minerals are commonly replaced by prehnite, laumontite, and, more rarely, sphalerite. Chlorite is the most abundant mineral in the surrounding dark green halos, partially to completely replacing clinopyroxene together with actinolite. Plagioclase is replaced by albite, prehnite, laumontite, and quartz, significantly more so than in the background basalts. Titanomagnetite is partially to completely replaced by titanite. Pyrite is a common accessory secondary phase.

The most highly developed alteration patch recovered in cores from Hole 1256D occurs in interval 312-1256D-174R-1, 91–109 cm, where ~15 cm of fine-grained basalt is completely replaced by greenschist alteration minerals in discrete mineralogic zones (Figs. F270, F271). The core of this alteration patch comprises euhedral, open space–filling quartz, epidote, and minor actinolite needles. Secondary porosity is filled by anhedral prehnite, and early prismatic quartz is replaced by laumontite. Secondary equant to tabular magnetite is common. The central core is surrounded by a ~10 mm chlorite-rich zone that grades into a distinctive yellow-green band intensively recrystallized to epidote, quartz, and actinolite. External to the epidote-rich zone, the igneous texture is preserved and light gray basalt is strongly recrystallized to laumontite, albite, actinolite, and chlorite. Large (~0.5 mm) pyrite grains occur along the diffuse boundary between the light gray basalt and the alteration patch. The patches are crosscut by several quartz-chlorite-pyrite veins and a late-stage quartz-laumontite-prehnite-chlorite vein.

Hydrothermal alteration patches are less common below 1280 mbsf in Hole 1256D. Below this level, patches are actinolite rich, with actinolite replacing clinopyroxene. Plagioclase is partially replaced by albite and chlorite, and titanomagnetite is recrystallized to subhedral titanite.

Alteration and mineralization associated with dike margins

As observed in the cores recovered from Expedition 309, dike-chilled margins are commonly the loci for brittle deformation, hydrothermal alteration, and sulfide mineralization (see “Structural geology”). Many of the dike margins recovered during Expedition 312 are highly fractured with the development of complex crosscutting vein relationships and associated alteration halos, indicating channeling of fluids along these contacts during cooling of the dikes (Fig. F272). Isolated slivers of host rock or xenocrystic fragments of differing grain size are commonly captured along the chilled margins; these fragments are completely recrystallized. Alteration generally occurs on both sides of the chilled margins, in the intruding dike and the host basalt. In some cases, hydrothermal alteration of the host rock preceded dike injection, as clearly indicated by the truncation of hydrothermal veins by dike margins. Chlorite and quartz are the most common components of veins associated with dike margins, although quartz is commonly replaced by laumontite or prehnite. Epidote, prehnite, laumontite, actinolite, titanite, pyrite, magnetite, and chalcopyrite are common vein fillings associated with margins. Rare sphalerite is present on some margins. Many veins have strongly developed chlorite-rich halos, with abundant secondary magnetite and sulfide grains, commonly concentrated directly along the chilled margin. Recrystallization of oxide phases and mobilization of iron along the dike margins is nicely exhibited in interval 312-1256D-175R-1, 43–46 cm, where spectacular flames of secondary magnetite erupt perpendicularly from the chilled margin into the marginal chloritic halo (Fig. F272). Larger grains of secondary magnetite are commonly acicular. Such iron oxide enrichments were not observed associated with the dike-chilled margins cored during Expedition 309.

The close relationship between dike margin deformation and hydrothermal alteration is clearly displayed in interval 312-1256D-179R-1, 5–9 cm (Fig. F273), in which an early actinolite-magnetite vein is offset by a number of small dike margin–parallel faults, whereas a later chlorite-quartz-titanite vein cuts across the margin without disruption.

Chronology of secondary mineral formation in the upper dikes of Hole 1256D

In the dikes drilled during Expedition 312, only a few unambiguous vein crosscutting relationships were observed, and a definitive petrogenetic sequence of mineral recrystallization and precipitation is, therefore, difficult to establish. However, the distribution of secondary minerals in the rare crosscutting veins provides useful constraints on the relative timing of alteration.

From synthesis of the available evidence, we propose the following alteration mineral associations for Hole 1256D sheeted dikes:

  • 1. Hornblende-rich amphibole (+ Ca-plagioclase),
  • 2. Actinolite + magnetite,
  • 3. Titanite + chlorite (+ quartz?),
  • 4. Quartz + epidote + sulfides,
  • 5. Prehnite,
  • 6. Laumontite,
  • 6. or 7. Calcite, and
  • 6. or 7. or 8. Anhydrite.

Group 1 results from the recrystallization of dike groundmass under amphibolite-facies conditions, and this occurs only relatively deep in the dikes, below ~1300 mbsf (Core 312-1256D-180R). Groups 2 and 3 occur within the dike groundmass at higher levels in the dike complex, and this alteration under greenschist-facies conditions overprints the Group 1 association in some samples from deeper in the dikes as the dike complex cooled. The formation of quartz, epidote, and sulfides in veins and patches results from more extensive hydrothermal fluid reaction under greenschist-facies conditions, and such assemblages overprint the background alteration. Precipitation of prehnite, laumontite, calcite, and anhydrite generally appears to be relatively late and probably occurs at slightly lower temperatures (100°–250°C). Veins of these minerals crosscut higher temperature veins and their associated halos. Rather than actual distinct stages of alteration, these various mineral associations reflect a continuous evolution of (decreasing) temperature and fluid composition through time.

Granoblastic dikes

A profound change in basalt texture and secondary mineralogy occurs in rocks deeper than 1348 mbsf (interval 312-1256D-190R-1, 0 cm, to 213R-1, 52 cm; 1348.1–1406.6 mbsf), and this portion of the sheeted dike complex will be described separately from the upper dikes. In hand specimen, the rocks appear to be moderately altered basalts macroscopically similar to the overlying dikes, but thin section observation reveals that significant proportions of these rocks are thoroughly recrystallized to microcrystalline, granular aggregates of secondary clinopyroxene, orthopyroxene, actinolitic hornblende, plagioclase, and subrounded blebs of magnetite and ilmenite. Development of the granoblastic assemblage is extremely heterogeneous, and only in rare examples are large areas of the igneous texture obliterated and completely recrystallized (Fig. F274) (e.g., Sample 312-1256D-194R-1, 36–37 cm). More commonly, the granoblastic dikes have only a minor component of completely recrystallized 0.5–1 mm patches included within zones in which there is only minor replacement of the original igneous texture by clinopyroxene and orthopyroxene. Subrounded, equant secondary magnetite is commonly the most visible indicator of partial recrystallization. In many samples, the primary variolitic to intergranular igneous texture is preserved by the in situ recrystallization of primary igneous minerals to subgrains comprising numerous equant, subrounded granules of clinopyroxene and orthopyroxene. Plagioclase laths are recrystallized to secondary plagioclase with subrounded grain boundaries. Larger euhedral plagioclase laths commonly have clear cores but inclusion-rich margins. Textural recrystallization of cryptocrystalline basalts is generally more advanced than neighboring coarser grained rocks.

The granoblastic assemblage can also occur as bands and veins. For example, Sample 312-1256D-198R-1, 46–49 cm, includes a >30 mm wide band of completely developed granular assemblage that is larger than the thin section (Fig. F275) and may be an isolated band or related to a vein. The granular assemblage can also be a major component of veins. Sample 312-1256D-207R-1, 10–15 cm, includes a vein made of large (up to 0.1 mm) euhedral orthopyroxene, chalcopyrite and magnetite, plagioclase, hornblende, minor quartz, and tiny (0.01 mm) “granular” clinopyroxene and orthopyroxene (Fig. F276). In Sample 312-1256D-209R-1, 8–10 cm, a vein composed of granular orthopyroxene, magnetite, minor clinopyroxene, and minor plagioclase is present. The adjacent rock (~0.5 mm) is completely recrystallized to orthopyroxene and magnetite, either as equigranular equant masses or mimicking the igneous "fanlike" clinopyroxene-plagioclase texture (Fig. F277).

To quantify the intensity of recrystallization to the granoblastic assemblage, a relative textural scale has been developed to classify basalts from the lower sheeted dikes. All rocks in the granoblastic dike zone can have isolated highly recrystallized patches, so the classification is applied to the background groundmass far away from patches and veins. Recrystallization begins with the growth of small, isolated micrometer-sized pyroxene granules replacing clinopyroxene and developing as small inclusions in clinopyroxene. With increased recrystallization there is “partial” replacement of the igneous texture and recrystallization to granular pyroxene (Fig. F278). Samples were classified as strongly recrystallized if all minerals are replaced but the igneous texture is only partially erased. In completely altered samples, the igneous texture is obliterated and overprinted by an equigranular granoblastic assemblage of secondary pyroxene, plagioclase, and magnetite. The heterogeneous development of the granoblastic texture in the sheeted dikes of Hole 1256D is well illustrated by the distribution of these features downhole (Fig. F279). Highly recrystallized patches and veins are present throughout the granoblastic dike zone, but many such features are present in rocks in which the groundmass is only partially recrystallized. Interestingly, the zone of most intense development of the granoblastic texture is from 1370 to 1397 mbsf, and the basalts directly overlying the gabbros are only partially to strongly recrystallized.

The relative timing of hydrothermal alteration and recrystallization of the granoblastic dikes is difficult to discern, and proof of hydrothermal alteration of these dikes preceding recrystallization is elusive. However, there is clear evidence of significant hydrothermal alteration postdating recrystallization and development of the granoblastic texture. An unambiguous crosscutting vein relationship between granular assemblages and hydrothermal veins is observed in Sample 312-1256D-209R-1, 8–10 cm (1396.58 mbsf), where a hornblende-plagioclase vein crosscuts and postdates the orthopyroxene-bearing vein described above (Fig. F280). In Sample 312-1256D-205R-1, 10–14 cm (1382.2 mbsf), the granular assemblage is strongly to completely developed but different extents of recrystallization and alteration styles are developed at various distances from the actinolitic hornblende-plagioclase-magnetite-quartz vein that crosscuts the sample. From the vein to the least altered wallrock, one successively observes a discontinuous (~1 mm) inner magnetite-rich halo, a central 1–3 mm light green actinolite-rich (50%) and dusty clinopyroxene-free alteration halo, and a dark gray outer alteration halo containing 20% actinolite and 20% dusty clinopyroxene/​actinolite, and in the outer, least altered zone secondary pyroxenes are not replaced by actinolite and dusty clinopyroxene/​actinolite (Fig. F281). Another element of temporal relationships is displayed in some veins where small equant clinopyroxene or orthopyroxene crystals are pseudomorphed by actinolitic-hornblende in the core of the vein but better preserved at the edge of the vein (Fig. F282) (e.g., Sample 312-1256D-197R-1, 3–6 cm; 1367.53 mbsf).

These observations constitute strong evidence for granular clinopyroxene-orthopyroxene-magnetite-plagioclase assemblages forming prior to the hydrothermal hornblende-actinolitic hornblende-plagioclase-quartz veins and the groundmass alteration occurring in the halos related to these veins. Within the upper dikes, actinolite replacing clinopyroxene is commonly riddled with tiny inclusions of magnetite. Rare secondary granoblastic clinopyroxenes have similar magnetite inclusions, and this may indicate an earlier phase of background hydrothermal alteration. Veins of orthopyroxene formed from hydrothermal solutions are certainly not common and possibly unknown in the oceanic crust, and these features may either represent recrystallization of preexisting hydrothermal veins with unusual composition or, perhaps more likely, thin magmatic dikelets.

Occurrence and alteration of oxide and sulfide minerals

Opaque minerals in the Expedition 312 sheeted dikes include igneous phases (titanomagnetite, ilmenite, and sulfide minerals) and secondary phases (magnetite, ilmenite, pyrite, chalcopyrite, sphalerite, and possible hematite). The distribution of opaque phases in Expedition 312 dikes can be divided into two zones on the basis of the alteration of igneous titanomagnetite: 1255.1–1314.5 mbsf (Cores 312-1256D-172R through 182R) and 1314.5–1406.6 mbsf (Cores 184R through 213R).

From 1255 to 1314 mbsf, titanomagnetite generally retains its igneous morphology, but in reflected light it appears corroded and may be rimmed and/or partly replaced by titanite. Exsolution lamellae of ilmenite are common in coarser grained samples and appear to be partly replaced by titanite. Finer grained titanomagnetite in microcrystalline and cryptocrystalline material at dike margins appears to be less altered than titanomagnetite in coarser grained rocks.

Secondary magnetite is common in this zone, where it occurs as tiny (up to several micrometer) blebs in partly altered clinopyroxene. The clinopyroxene appears dusty, slightly brownish, and corroded and is commonly rimmed or partly replaced by actinolite, which may contain similar blebs of secondary magnetite.

Pyrite and chalcopyrite are common from 1255 to 1314 mbsf. These phases occur in minor amounts as recrystallized igneous sulfide globules, either as inclusions in plagioclase or in interstitial areas. Secondary pyrite and lesser chalcopyrite are commonly disseminated in the rocks, replacing silicates and in interstitial areas. These phases are also common in veins, mostly intergrown with quartz, although chlorite and prehnite are also associated (apparently as earlier and later generations of vein minerals, respectively). One quartz-bearing vein contains small (a few micrometers) rounded grains of sphalerite. Veins at chilled dike margins in the upper zone commonly contain pyrite, chalcopyrite, and magnetite in various combinations. Pyrite and chalcopyrite are typically associated with quartz in these veins, but magnetite may occur with quartz or chlorite.

Beginning at ~1314 mbsf (Core 312-1256D-184R) and continuing at least through 1369 mbsf (Core 198R), the alteration of igneous titanomagnetite changes. Primary titanomagnetite is completely recrystallized to irregular to globular grains of magnetite ± ilmenite ± small amounts of titanite. A mineral that appears slightly bluish and has relatively high reflectivity also occurs in small amounts within recrystallized magnetite and may be hematite. Globular to rounded grains of secondary magnetite are commonly included within actinolite replacing clinopyroxene in this interval but are coarser grained than in the overlying rocks. Secondary clinopyroxene appears deeper within this lower interval of the hole, and this secondary clinopyroxene commonly contains several micrometer-sized blebs of secondary magnetite, similar to those within actinolite. Secondary sulfides from below 1314 mbsf are generally similar to those in the immediately overlying rocks, but occurrences that are clearly recrystallized igneous sulfide globules were not recognized.

Plutonic section

The plutonic section (interval 312-1256D-213R-1, 52 cm, through 234R-1, 33 cm; 1406.62–1507.1 mbsf) comprises dark gray to dark green-gray gabbroic rocks, leucocratic oxide diorites and trondhjemites, and dike screens. All rocks are moderately to completely altered. Most samples display abundant signs of hydrothermal alteration, visible in hand specimen. Whole-rock alteration of gabbros and associated plutonic rocks is quantified in the plutonic rock alteration log (see PLUTLOG.XLS in “Supplementary material”), whereas veins were recorded in the standard vein log (see 312VEIN.XLS in “Supplementary material”). Alteration of basalts from the dike screens was recorded in the standard alteration log (see 312ALT.XLS in “Supplementary material”). The coarser grain size of many samples from the plutonic section allowed the extent of alteration of principal igneous minerals to be estimated following calibration by thin section observations. Intensity of alteration is displayed for whole rocks and mineral by mineral in Figure F283. Distribution and abundance of veins are shown in Figures F259 and F260 and summarized in Table T42.

Alteration of gabbros

The extent of gabbro alteration is principally dependent on the intensity of replacement of plagioclase and clinopyroxene by secondary minerals. In hand specimen, rocks are dark gray to dark gray-green, with a greenish hue signifying more extensive replacement of clinopyroxene by actinolitic hornblende (Fig. F284). Plagioclase, when altered, generally appears whiter than igneous feldspar, with the margins of some highly altered crystals blue-green where replaced by intergrown secondary feldspar and actinolite + chlorite. Disseminated pyrite is common throughout the plutonic section. Pyrrhotite with chalcopyrite and magnetite occurs as centimeter-scale clots in Section 312-1256D-223R-3, 6 cm, and coarse-grained (1–5 mm) disseminated groundmass pyrrhotite is common from Section 231R-4 (~1491.9 mbsf) to the bottom of the hole.

The intensity of alteration is strongly dependent on the grain size of the rock. Regions of gabbro with large (~10–15 mm) ophitic clinopyroxenes tightly intergrown with plagioclase are much less altered than the irregular coarser grained zones in between, which tend to be highly to completely altered with obvious amphibole and white secondary plagioclase clearly visible in hand specimen. Large irregular patches of pegmatitic gabbro are similarly altered to amphibole and secondary plagioclase. In general, leucocratic rocks (e.g., Units 1256D-82, 90b, 90c, and 90f) are more altered than the host gabbros or dike, although this may reflect the relative abundance of such rock types near intrusive margins and the narrow (~5–15 mm) width of many leucocratic intrusions. These rocks are commonly altered to amphibole, secondary plagioclase, chlorite, epidote, prehnite, and titanite. Recrystallization of quartz is difficult to quantify. The host rock around such leucocratic intrusions is commonly more intensively altered, with 2–15 mm dark green amphibole and chlorite-rich halos that exhibit common 1–2 mm black irregular patches in the groundmass and intense to complete replacement of clinopyroxene by actinolitic hornblende and of plagioclase by secondary feldspar, actinolitic hornblende, prehnite, and epidote.

The external intrusive margins at the top of the plutonic section and the boundaries with the upper dike screen are the loci for extensive hydrothermal alteration. The gabbro directly below the granoblastic dikes (Section 312-1256D-213R-1, 52 cm) is highly to completely altered, with extensive replacement of clinopyroxene by green actinolitic hornblende and of plagioclase by secondary feldspar, zeolite, and abundant 5–10 mm sized clots of epidote. A 2 mm wide greenish actinolite-rich halo is developed along a 0.5 mm magnetite vein that delineates the boundary between the sheeted dikes and upper gabbros in Hole 1256D. The gabbro directly below the margin is completely recrystallized with visible grain size reduction due to intense hydrothermal alteration. Alteration is high to complete throughout Section 312-1256D-214R-1, with numerous pieces totally replaced by secondary minerals, including abundant epidote (Fig. F284). Similarly, the top of the lower gabbro (Gabbro 2; Section 312-1256D-230R-1, 15 cm) (Fig. F284) is a highly to completely altered heterogeneous dark gray to dark greenish gray gabbro (Unit 1256D-91a). Here the upper boundary of the gabbro intrusion is completely altered to secondary feldspar + epidote + amphibole + chlorite. The host basalt dike (Unit 1256D-90a) is baked and has a 5 mm wide very dark gray alteration halo along the vertical margin with the gabbro intrusion.

In thin section, clinopyroxene within gabbros appears highly to completely pseudomorphed by actinolitic hornblende with common inclusions of magnetite (Fig. F285); minor chlorite may be present. Actinolitic hornblende commonly occurs with an irregular matted texture where pyroxene is completely replaced. Where there is incomplete replacement, clinopyroxene cores are partially replaced by dusty/​corroded clinopyroxene and completely surrounded by actinolitic-hornblende rims. Along cleavage planes and cracks, clinopyroxene is intensely altered to amphibole. Plagioclase alteration is generally less advanced than clinopyroxene, and it is partially altered to secondary plagioclase (albite), actinolitic hornblende, chlorite, prehnite, laumontite, and epidote (Fig. F286). Pumpellyite was tentatively identified replacing plagioclase in Sample 312-1256D-217R-1, 64–69 cm (1422.24 mbsf). If confirmed, this would be only the second documented occurrence of pumpellyite from in situ ocean crust, with pumpellyite being first identified in the lower sheeted dikes of Hole 504B (Alt, Kinoshita, Stokking, et al., 1993). From Core 312-1256D-214R to Core 232R (~1411–1495 mbsf), most plagioclase laths retain their igneous habit and still exhibit magmatic zoning, although many crystals appear gray because of the very large number of tiny inclusions aligned in strands parallel to twin planes within the crystals. This feature is more pronounced with depth and also occurs within the dike screens and could therefore indicate recrystallization at high temperatures. It should be noted that although gray plagioclase is recorded in thin section descriptions as unaltered it may be significantly recrystallized and have secondary compositions. Confirmation of the origin of this plagioclase will require postcruise investigation.

Olivine is present in small amounts (1%–3%) in most plutonic rocks but is abundant in Section 312-1256D-223R-2 (~1451 mbsf) where it originally made up ~20% of primary minerals. In hand specimen, altered olivine appears as dark green chloritic blebs commonly with tiny bright red inclusions. In thin section, olivine is partially to completely altered to pleochroic brown-green phyllosilicates (undetermined), talc, and magnetite with outer rims of pale blue-green amphibole and minor chlorite (Fig. F287). Iron oxyhydroxides and/or hematite within altered olivines impart a red color in hand specimen (e.g., Sample 312-1256D-220R-1, 8–10 cm).

Ortho-amphibole has been tentatively identified in Sections 312-1256D-213R-1 and 214R-1 (Fig. F288). It occurs in minor amounts as colorless, high-relief, elongated (width = 0.1–0.3 mm; length = 1–5 mm) crystals. It shows cleavage perpendicular to its elongation, low birefringence (first-order gray), and parallel extinction. Ortho-amphibole overprints primary magmatic phases and therefore postdates them. However, the timing relation of ortho-amphibole with respect to other secondary minerals remains unclear.

Titanomagnetite exhibits ilmenite exsolution lamellae and is typically partly to highly replaced by titanite. Recrystallized igneous sulfide globules are common in interstitial areas and as inclusions in plagioclase and rarely in actinolitic hornblende pseudomorphs after clinopyroxene. These consist of irregular to rounded globules, up to 500 µm, of pyrite ± chalopyrite, local pyrrhotite, and rare millerite. Secondary pyrite, trace chalcopyrite, and rare pyrrhotite are present locally in interstitial areas intergrown with actinolite and rarely in veins associated with amphibole, quartz, and magnetite.

Gabbros contain fewer veins than the overlying sheeted dikes with, on average, 10 veins/m of recovered core compared to ~35 veins/m in the sheeted dikes (Figs. F259, F260; Table T42). The earliest generation of veins is difficult to discern and occurs as diffuse 1–2 mm wide actinolitic hornblende halos with no discrete vein filling. These veins are crosscut by discrete 0.5–1 mm actinolitic hornblende veins, also with alteration halos. In zones of intense hydrothermal alteration, actinolite veins are cut by epidote, quartz, and prehnite veins with intensely altered chloritic margins (Fig. F289). These relatively early vein sets are crosscut by numerous 0.5 mm chlorite and later ~1 mm quartz-chorite veins, with 2–5 mm green alteration halos.

Upper dike screen and basaltic rocks below 1494.9 mbsf

The lower boundary of Gabbro 1 was not recovered in Core 312-1256-224R, and the classification of the basalts in this interval as a dike screen (interval 225R-1, 0 cm, through 230R-1, 15 cm; 1458.9–1483.1 mbsf) is based principally on evidence for textural recrystallization. The lower boundary of the dike screen with Gabbro 2 is clearly intrusive with stoped clasts of partially resorbed basalt within the medium-grained gabbro of Unit 1256D-91a. These basalts (Unit 1256D-90a; interval 312-1256D-232R-2, 97 cm, through 234R-1, 33 cm; 1494.9–1507.1 mbsf) are dark gray fine-grained to cryptocrystalline basaltic dikes that are highly to completely altered with common disseminated pyrite. In hand specimen, they have a speckled or mottled appearance, suggesting granoblastic recrystallization, although an igneous texture is apparent in places. In thin section, the basalts are partially recrystallized to granular assemblages with smooth annealed grain boundaries, perhaps suggesting reheating by the gabbro body intruded above (Fig. F290A). The igneous texture of the dikes is maintained by plagioclase, but clinopyroxene is partially to completely recrystallized to subrounded grains. Titanomagnetite is partially recrystallized to subrounded grains, but many grains have more angular shapes compared to grains in the granoblastic dikes overlying the upper gabbro (Gabbro 1). Clinopyroxene is replaced by dusty clinopyroxene or actinolitic hornblende + magnetite. Plagioclase appears mostly fresh, although grain edges are smooth and rounded. It commonly has a light gray appearance due to cores of crystals being host to innumerable tiny inclusions, although the rims of plagioclase are commonly inclusion free. Alteration is patchy with some zones containing abundant granular orthopyroxene, but other regions are dominated by actinolitic hornblende or dusty, corroded clinopyroxene. Orthopyroxene occurs as small equant, inclusion-free grains, the rims of which are commonly altered to actinolitic hornblende and chlorite (Fig. F291). Coexisting clinopyroxene is riddled with magnetite crystals.

The basalts contain numerous veins (~15–30 veins/m) (Figs. F259, F260; Table T42) of a number of different generations, most with halos. The first generation of veins is wispy <0.1 mm actinolitic-hornblende veins, and their location is signaled by the presence of 1–2 mm diffuse halos. These veins are crosscut by later actinolitic-hornblende veins with more discernible fillings, and both generations have highly recrystallized actinolite-rich halos. Later chlorite-actinolite, quartz-chlorite, and 0.5–2 mm braided quartz veins with 1–2 mm chloritic margins cut across the earlier vein generations (Fig. F292).

Recrystallized basaltic dikes are intruded by a number of thin (<20 mm) mafic and felsic dikes (Units 1256D-90b through 90f). Units 1256D-90b, 90c, and 90f are trondhjemite and quartz-rich oxide diorite dikelets that are highly altered to secondary plagioclase, actinolitic-hornblende, chlorite, titanite, epidote, prehnite, sulfides, and minor calcite. The margins of these dikelets are commonly diffuse, and the host wallrock is more highly altered. Gabbroic dikelets (Units 1256D-90d and 90e) are similarly highly altered to secondary plagioclase, actinolitic-hornblende, epidote, quartz, prehnite, and magnetite with >2 mm dark green halos in the surrounding wallrock. Crosscutting actinolite + chlorite veins with 2 mm light gray halos indicate significant hydrothermal alteration postdating the intrusion of these dikes.

The most altered sample of the plutonic section of Hole 1256D occurs in the dike screen (Sample 312-1256D-227R-1, 87–91 cm) where a trondhjemite is composed dominantly of epidote (average = 40 vol%) and quartz, chlorite, titanite, albite and laumontite. Magmatic phases and textures are locally completely erased by this hydrothermal alteration (Fig. F284).

Interestingly, the main secondary mineral in the fine-grained basalt in contact with the gabbro at the top of Gabbro 2 (Sample 312-1256D-230R-1, 49–54 cm; 1483.49 mbsf) is a pleochroic dark green to olive-green mineral with low birefringence that partly or completely replaces subrounded granular orthopyroxene. This mineral is tentatively identified as a smectite-rich phyllosilicate. If confirmed by postcruise analyses, this would suggest that parts of the plutonic section have been altered at relatively low temperatures (<200°C).

The lower portion of Gabbro 2 (interval 312-1256D-232R-2, 97 cm, through 234R-1, 33 cm; 1494.9–1507.1 mbsf) is complicated and comprises an intrusive contact between the overlying gabbro (Unit 1256D-93) and fine-grained basaltic dikes (Unit 1256D-94). Rounded, partially resorbed basaltic xenoliths are observed in the lower sections of Gabbro 2 (see “Igneous petrology”). Unit 1256D-94 is composed of highly recrystallized orthopyroxene-bearing basalts that display textures and alteration patterns similar to those in basalts of the upper dike screen. However, there is abundant coarse-grained (1–5 mm) pyrrhotite in these basalts. The contact with the gabbro is sharp and sutured, but 1–5 mm diffuse gabbro stringers intrude the fine-grained basalt in Section 312-1256D-233R-1. These gabbro dikelets have ~5 mm dark green halos.

The contact between Units 1256D-94 and 95 is not intact, but an oriented piece (interval 312-1256D-234R-1, 20–29 cm) indicates that Unit 1256D-95 was recovered in situ and is not merely rubble. This basalt is dark greenish gray with abundant disseminated pyrrhotite and includes several 1 mm quartz + chlorite veins. Although highly altered, the rock does not have a granoblastic texture, but instead it is reminiscent of dikes from the lowermost upper dikes (e.g., Core 312-1256D-180R; 1300 mbsf), in which clinopyroxene is altered to dusty clinopyroxene-actinolite, actinolite, chlorite, and secondary plagioclase while maintaining the primary intergranular igneous texture (Fig. F290B). The absence of granoblastic texture suggests that this basalt was intruded into the lower dike screen and possibly the overlying gabbros.

Structural geology

Cores recovered from Hole 1256D between 752.0 and 1507.5 mbsf contain structures of magmatic, alteration, and deformational origins (Fig. F293). Structures described during part of Expedition 309 and Expedition 312 between 1060 and 1406 mbsf are solely within the sheeted dike complex. Below 1406 mbsf, structural features are mostly within gabbroic rocks. Features described during Expedition 312 thus differ from those described during Leg 206 and much of Expedition 309 because many features described above 1060 mbsf are syn- and posteruptive structures developed in lavas.

Techniques and methods used to describe and document structures are discussed in “Structural geology” in the “Methods” chapter. Observations were recorded on the structural description forms (see “Supplementary material”) and in the structural log (see STRUCTUR.XLS in “Supplementary material”). In the following sections, the nature of the structures is described, followed by a discussion of the distribution and orientations of and relationships between structures and then a short summary.

Some important observations and interpretations from Expedition 312 include the following:

  • Preferred orientations of planar and subplanar structures, variation in structural morphology, and crosscutting relationships are evidence for close linkages between intrusive processes, fluid flow, and brittle deformation.
  • Structures within the granoblastic dikes are more gently dipping than in the overlying upper sheeted dikes, indicating that the structures were affected by the thermal and/or stress field imposed by the underlying gabbro.
  • The gabbro is a moderately dipping body with structures indicative of both porous and fracture (dike) melt-flow.
  • The gabbro exhibits little to no evidence for shear strain, and local brittle structures formed during alteration at greenschist to amphibolite conditions.

Magmatic flow structures within dikes

Magmatic flow structures exhibit evidence of particle rotation or plastic strain imposed by the flow of viscous magma (Passchier and Trouw, 1996; Dick et al., 1992). Most magmatic flow structures are located near or at dike margins, discussed below. Within dikes, locally aligned plagioclase laths define trachytic textures (e.g., interval 312-1256D-192R-1, 11–13 cm) (Fig. F294A; see “Igneous petrology”). Alteration patches within dikes may be related to deformed primary porosity, amygdules, or inherited textural inhomogeneities that were strained by viscous flow during and/or shortly after emplacement (Bach and Irber, 1998) (see “Alteration-related structures,” below, and “Alteration”).

Chilled margin magmatic flow structures include the preferred orientation of plagioclase laths that are parallel to the dike margin (e.g., interval 312-1256D-176R-2, 0–2 cm) (Fig. F294B), folding and imbrication of flow bands comprising disseminated oxides in cryptocrystalline-to-glassy material (rheomorphs) (e.g., interval 176R-2, 0–2 cm) (Fig. F294C, F294D) and stretched spherulites (e.g., intervals 176R-2, 0–2 cm, and 187R-1, 75–77 cm) (Fig. F294E, F294F; see “Igneous petrology”).

Magmatic flow structures within gabbros

Similar to dikes, gabbros contain grain-scale magmatic flow structures, such as shape-preferred orientation of elongate crystals near contacts. At the >1 cm scale, gabbros contain a variety of structures and fabrics indicative of magmatic flow. Shallowly inclined bands that have different grain size and/or composition are common in gabbros (e.g., interval 312-1256D-232R-1, 70–99 cm, and Section 232R-3 [Piece 1]) (Fig. F295A, F295B). These bands are hereafter referred to as textural bands and color bands. In one place (Section 312-1256D-223R-3 [Piece 1]) (Fig. F295C), a weakly developed magmatic shear zone crosscuts textural banding. Much of the original structure of the shear zone has been overprinted by recrystallized plagioclase (Fig. F295D). In places, lineation and foliation are defined by alignment of plagioclase (e.g., interval 312-1256D-232R-1, 70–99 cm). Although these structures and fabrics are recrystallized, there is no evidence for crystal-plastic deformation and we interpret them as forming via magmatic flow.

Another magmatic structure commonly present in gabbros is local domains of contrasting texture and/or composition distributed within more heterogenous gabbro (e.g., Sections 312-1256D-232R-3 [Piece 1] and 220R-1 [Piece 3]; intervals 223R-2, 36–55 cm, and 214R-3, 5–13 cm) (Figs. F295C, F296). These magmatic patches are commonly leucocratic domains of coarser grained, plagioclase-rich aggregates within otherwise melanocratic gabbro. Melanocratic gabbro is composed of plagioclase-clinopyroxene aggregates with ophitic texture (see “Igneous petrology”). Magmatic patches do not transect the split core surface and hence are not considered bands. Magmatic patches have a range of different morphologies, from interconnected, amoeboid with irregular boundaries to elongate. Apparent aspect ratios of patches measured on the cut surface of the core are mostly equant, although elongate patches are also present (e.g., interval 312-1256D-214R-2, 61–70 cm) (see Fig. F297). Alteration and recrystallization largely overprint these patches, preventing simple interpretations of flow structures within them (Fig. F295D, F296; see “Alteration”). Less thoroughly recrystallized patches have been described in cores from the Mid-Atlantic Ridge and proposed to be the product of melt flow through gabbro (Blackman, Ildefonse, John, Ohara, Miller, MacLeod, et al., 2006).

Intrusive contacts

Gabbro-dike and intragabbro contacts

Contacts between gabbros and dikes, and between units within the plutonic section, have several structural and alteration properties of note (see “Alteration” and “Igneous petrology”). The uppermost contact of Gabbro 1 (Fig. F298A–F298F) exhibits a slight increase in grain size within <2 cm of the contact. Pervasive alteration and recrystallization extend from the contact into the overlying granoblastic dikes (Section 312-1256D-213R-1 [Pieces 12 and 13]). Adjacent to the gabbro/dike contact, a dense network of veins and fractures, now largely recrystallized, create a pseudocataclastic texture. Perhaps because of the recrystallization, there is no identified evidence for shear strain or magmatic flow at the contact. Thin leucocratic intrusions into the overlying dikes are well-defined dikelets with no internal flow fabrics (Fig. F298A).

The contact between Gabbro 2 and the dike screen contrasts with the contact between Gabbro 1 and the overlying granoblastic dikes (Section 312-1256D-230R-1 [Piece 5]) (Fig. F298G, F298H). The area in the gabbro in contact with the overlying dike screen has a millimeter-scale shape-preferred orientation of plagioclase that is parallel to the contact (Fig. F298H). At a larger scale, magmatic structures are related to xenoliths of basalt within the gabbro (Section 312-1256D-230R-1 [Piece 8]) (Fig. F298I–F298L). In places, xenoliths have a weak shape-preferred orientation (Fig. F298I), although they have no clear indication of viscous or plastic flow. The boundaries of each xenolith have a 5 mm area of reduced grain size, extensive recrystallization, and weak oxide- and/or plagioclase-defined flow structures inside the host gabbros (Fig. F298J–F298L). One of the xenoliths has boundaries that are discordant with weak alignment in the surrounding gabbro (Fig. F298J). Thus, there is evidence that the basaltic xenoliths are the product of the intrusion of the gabbro, but contacts between gabbro and basaltic xenoliths are modified by a combination of melting, recrystallization, and alteration.

Chilled dike margins

Most intrusive contacts within the sheeted dikes have planar or subplanar chilled margins. Chilled margins are distinguished by changes from micro- to cryptocrystalline grain sizes to, in places, glassy domains within a few centimeters of the contact with fine-grained basalt (e.g., Sections 312-1256D-175R-1 [Piece 11] and 179R-1 [Piece 2]) (Fig. F299). The grain-size variation is part of a more general banding of texture and grain size. The band at the contact is a baked margin and is in many pieces replaced and/or cut by chlorite, sulfide, and oxide veins. This baked zone and the related alteration extend >1 mm into the wallrock. The band adjacent to the baked zone, within the chilled margin, is a spherulite and crystallite band that is white in hand sample but transparent in thin section. Adjacent to the band of spherulites is a band of unknown composition that is dark in hand specimen but dusty to opaque in thin section (Fig. F299C, F299D, F299E).

Chilled margin structures are overprinted by later processes such as veining and alteration, and this makes it difficult to distinguish the primary properties of chilled margins. For example, the dark band in Section 312-1256D-179R-1 (Piece 2) could be an alteration feature or a chill related to a second intrusion. In one piece (interval 312-1256D-175R-1, 43–58 cm), a dark gray band has an en echelon geometry (Fig. F299H). This arrangement is consistent with the interpretation that the dark bands are microintrusions of basaltic material. The structures discussed in the above section, “Magmatic flow structures within dikes,” indicate that chilled margin structures are largely due to the flow of viscous magma. In some places, where flow structures are not present, spherulitic texture is reminiscent of grains that “settled” in a viscous material (Fig. F299F, F299G).

Also obscured by alteration, there are places (e.g., intervals 312-1256D-187R-1, 75–77 cm, and 194R-1, 8–14 cm) where chilled margins contain inclusions of “exotic” material (Fig. F300). Elongate inclusions have their long axes oriented parallel to the chilled margin (e.g., intervals 312-1256D-194R-1, 8–14 cm, and 187R-1, 75–77 cm) (Fig. F300A, F300B). In places, inclusions are not directly correlative with the adjacent wallrock, yet are likely of basaltic protolith. In some pieces, the clasts are so intensely altered that it is unclear if they are merely alteration patches rather than detached clasts of wall rock.

Magmatic veins in sheeted dikes

A possible flow apophysis or magmatic vein (or dikelet) is present in the recovered cores (Section 312-1256D-186R-1 [Pieces 11 and 12]) (Fig. F300C–F300E). The vein is irregular, 4 mm wide, and composed of pervasively altered glassy basalt. The magmatic vein contains an inclusion that correlates with the wallrock. Flow banding around the inclusion demonstrates that it was entrained during a period of viscous melt flow. Another interesting feature of the magmatic vein is the presence of fractures throughout that do not everywhere clearly cut the wallrock. In places, the fractures are filled with small amounts of actinolite, resulting in a microbreccia domain developed in the dikelet.

Brittle structures around dike margins

Many dike margins and the adjacent host rocks contain fractures, veins, and breccias. These dike margins are referred to here as brecciated. Brecciated dike margins have the same properties as unbrecciated chilled margins, such as bands with differing textures parallel to the margins (e.g., Sections 312-1256D-175R-1 [Pieces 11 and 12] and 176R-2 [Pieces 1 and 2]). Within well-developed veins and domains of breccia, chlorite, amphibole, oxides, and sulfides form a matrix surrounding angular clasts of host rock and margin material (Figs. F301, F302). Veins occur most commonly at dike margins and extend into adjacent wallrock, forming networks (Fig. F301).

One of the noteworthy properties of brecciated dike margins is the close relationship between brittle features and magmatic structure. For example, magmatic flow fabrics are parallel to and infolded around veins of breccia (Fig. F294C, F294D). Elsewhere, veins offset the margins but also control changes in the alteration pattern (Fig. F301). In places, breccia domains and veins in the host rock are cut by cryptocrystalline basalt (Fig. F302). Furthermore, veins are locally offset in small steps across chilled margins (Fig. F302C). These close spatial, geometric, and crosscutting relationships between breccias, veins, alteration, and magmatic structures suggest a genetic linkage.

Deformation mechanisms

Basaltic and gabbroic rocks recovered from Hole 1256D during Expedition 312 are essentially unstrained. There is little to no evidence for crystal-plastic or extensive cataclastic strain, and magmatic flow structures are subtle except when adjacent to contacts (see “Magmatic flow structures within dikes,” above). Despite the low bulk strain of the rocks, there are grain-scale structures related to alteration and magmatic processes, which are best expressed in the plagioclase grains.

In dikes, glomerocrysts of plagioclase are in places intensely fractured, possibly related to cooling and contraction (e.g., interval 312-1256D-174R-1, 130–134 cm) (Fig. F303A). Within the granoblastic dikes, plagioclase laths are locally bent (e.g., interval 312-1256D-198R-1, 46–49 cm) (Fig. F303B), accommodating modest strain by intracrystalline deformation mechanisms (e.g., Passchier and Trouw, 1996).

In gabbros, many structures are overprinted by recrystallization. Fractures are effectively annealed by secondary plagioclase (e.g., intervals 312-1256D-214R-1, 15–17 cm, and 215R-1, 84–88 cm) (Fig. F303C, F303D). Where not recrystallized, plagioclase grains are intensely fractured (e.g., interval 312-1256D-214R-2, 0–6 cm) (Fig. F303E). In some places, the intense fracturing gives rise to local cataclastic zones (e.g., interval 312-1256D-214R-1, 121–124 cm) (Fig. F303F).

Distribution of fracturing and cataclasis is very heterogeneous. Brittle structures are commonly proximal to alteration patches and amphibole grains. There is a lack of annealing of some fractures, and actinolite needles are present in some cataclastic zones. These observations suggest that this generation of brittle grain-scale deformation occurred at greenschist conditions after the recrystallization of other plagioclase grains.

Fractures with no mineralization (including joints)

A variety of fractures are described in cores recovered during Expedition 312 (Fig. F304). Drilling-induced fractures are in most places shallowly dipping and create "saddles" toward the center of the section (see “Veins and joints” in “Structural geology” in “Expedition 309”; see also Fig. F145) (e.g., intervals 312-1256D-187R-1, 56–85 cm, and 194R-1, 0–59 cm) (Fig. F304). In some sections, the shallowly dipping fractures have associated, possibly conjugate, high-angle fractures nearby (Fig. F304). Submillimeter white fractures protrude from the edges of many pieces and taper toward the center (Fig. F304). These fractures have no mineral filling, although some comminuted dike rock causes the white color. Most shallowly dipping fractures are considered to be drilling related, although they may nucleate on preexisting features. In cores with particularly low recovery, the spacing of drilling-induced fractures is roughly every 2 cm, more or less the size of the smallest whole-round pieces (Fig. F304E, F304F).

Alteration-related structures

Alteration patches

Alteration patches are domains with abundant secondary minerals such as chlorite or actinolite (Fig. F293; see “Alteration”). Apparent aspect ratios of relatively well defined ovoidal patches were measured on the cut surface of the archive half. In the sheeted dike complex, elongate alteration patches may be indicative of a relict magmatic flow fabric (e.g., interval 312-1256D-174R-1, 70–73 cm) (Fig. F305). It is not known exactly what phases the secondary minerals replace, but it has been suggested that they replaced glass or highly fractionated crystallized melt or that they filled pore space (Alt, Kinoshita, Stokking, et al., 1993; Bach and Irber, 1998) (see “Alteration”). Some of the patches are elliptical in shape (Fig. F305). There is no evidence to suggest that patches were strained after their formation. Other patches are not clearly related to any primary feature (Fig. F305). These patches may result from alteration by fluids that flowed through the relatively undeformed basalt. In some pieces, veins and alteration patches coexist, with veins cutting the center of spherical patches and passing along the center of elongate patches (e.g., interval 312-1256D-176R-1, 6–18 cm) (Fig. F305E–F305G).

Alteration patches in gabbros are composed mostly of actinolite (see “Alteration”), and their shapes are mostly equant (Fig. F297). Most alteration patches have an apparent aspect ratio (R) of 1, and the number of patches with R > 1 decreases toward R > 4.5 (Fig. F297E–F297G). Alteration patches in the sheeted dikes do not show the same trend.

Veins

Veins are the most common structure in recovered cores. Most of the inspected pieces contain at least one vein. Veins range in width, morphology, and composition and reflect the dynamics of hydrothermal fluid flow through the crust (see “Alteration”). Vein widths are generally <1 mm (most are ~0.1 mm wide), with a maximum width of 5 mm in interval 312-1256D-186R-1, 53–57 cm. In a few pieces, veins—particularly quartz sulfide–bearing veins—are >1 mm wide. Many veins have alteration halos commonly >1 mm wide (Fig. F306) (e.g., intervals 312-1256D-173R-1, 107–128 cm, and 180R-1, 0–4 cm). The color of the alteration halos reflects secondary mineral contents. Many veins, particularly those with well-defined walls, do not have halos (Fig. F306).

Veins have a range of morphologies including planar, curved, irregular, and anastomosing (Figs. F307, F308). In some sets, veins have splays (Figs. F307, F308) and intersect others with Y- and T-shaped morphologies (Figs. F307, F308). These intersections can be complex, with changes in mineralogy along the length of the vein. Furthermore, many veins that apparently merge in hand sample have crosscutting relationships in thin section (Figs. F307, F308). Crosscutting relationships include the merged veins but can also include displacements at vein intersections (Fig. F307).

Vein fill changes from chlorite to actinolite and to hornblende with depth in Hole 1256D (see “Alteration”). Chlorite-rich veins commonly show compositional zoning with chlorite against the wallrock and quartz in the center (Fig. F306) (e.g., interval 312-1256D-181R-1, 27–30 cm). Sulfides in the veins are texturally late and in many instances in the vein center (Fig. F306) (e.g., interval 312-1256D-176R-1, 92–94 cm).

The zonation in mineralogy from chlorite on the vein walls to quartz in the center is interpreted to signify that the veins are antitaxial, wherein grains grow from solution in the vein center during incremental opening. The vein minerals do not replace the wallrock, further supporting the antitaxial interpretation. Some models for antitaxial vein opening suggest that fibrous and elongate minerals can track the opening history (Ramsay and Huber, 1983). In some veins from Hole 1256D, slightly elongate grains of quartz and cleavage planes of chlorite are interpreted to have tracked the vein opening direction (Fig. F306F). Such structures are rare but indicate that vein opening was slightly noncoaxial. In the case of shear veins, vein opening was strongly oblique (see “Shear veins”).

Shear veins

Unlike Expedition 309, where local cataclastic rocks were observed, faults and fault rocks were not recovered during Expedition 312. However, several shear veins were recovered. Shear veins exhibit evidence for noncoaxial displacement either during or after their opening (Ramsay and Huber, 1983). Slickenfibers, also called mineral lineations, are present on some surfaces (Fig. F309). Such planar surfaces commonly have steps in the slickenfibers (Fig. F309B), and the standard interpretation is that displacement was in the stepping-down direction (Ramsay and Huber, 1983). Only two of the five shear veins recovered exhibit sense of shear indicators, one sinistral and one normal (Sections 312-1256D-173R-1 [Piece 12] and 202R-1 [Piece 2]). Shear veins in thin section are filled with blocky quartz, minor carbonate, and lenses of chlorite, actinolite, and epidote (Fig. F309D, F309E). Secondary minerals apparently predate much of the vein opening and provide passive markers for the opening history. In contrast, quartz grains track the vein opening and surround inclusions of secondary minerals. Locally, shear veins exhibit strain in the form of patchy and/or undulose extinction, stretched quartz, and domains of grain-size reduction (Fig. F309F). Shear veins are also present in a specific interval of the recovered gabbro (interval 312-1256D-231R-2, 77–85 cm) (Fig. F309G).

Downhole distribution of structures at Site 1256

Structures are grouped into fractures, veins, shear veins, breccias, and chilled margins/igneous contacts (Fig. F293). This grouping combines structures of similar types rather than similar mechanisms. For example:

  • Breccias above 1000 mbsf are in most places related to the deformation imposed by emplacement, cooling, and burial of lavas.
  • Breccias in the transition zone between 1000 and 1060 mbsf are formed by a combination of volcanic, deformation, and dike-emplacement processes.
  • Breccias below 1060 mbsf are near dike margins and probably relate to dike emplacement.

Other factors influence the downhole distribution of structures (Fig. F293). Alteration patches were only logged during Expedition 312. During Leg 206 and Expedition 309, logged features were measured only in oriented pieces, whereas features were logged in both oriented and unoriented pieces during Expedition 312 (see “Structural geology” in the “Methods” chapter).

Fractures are clustered into groups scattered throughout the hole. Below 1280 mbsf, fracture density decreases (Fig. F310). Interestingly, fracture intensity is only greater than high in three intervals in core recovered from Hole 1256D: near the top of the sheet and massive flows, near the top of the transition zone, and in the upper sheeted dike complex (Fig. F310). Although poor recovery may be related to the spatial distribution of fractures, it does not appear to be related to fracture intensity (cf. Fig. F268).

Veins are distributed throughout the hole with gaps in several intervals perhaps related to recovery (Fig. F293). Shear veins are also distributed throughout the hole, although they decrease in abundance below 1300 mbsf where dikes are more prominent and recovery was low. The distribution of alteration patches is widely scattered, although distribution may be an artifact of recovery. Igneous contacts are distributed between 765 and 1500 mbsf and are apparently absent from depths above 765 mbsf. The interval containing the most igneous contacts is between 1090 and 1290 mbsf, in the upper sheeted dike complex. Another interval with numerous igneous contacts is between 1450 and 1500 mbsf within gabbroic rocks. Breccias are sparsely distributed between 348 and 1275 mbsf (Fig. F293).

Downhole orientation of structures

The data set of structural orientations from Expedition 312 reveals preferred orientations grouped both by unit and by type of structure (Figs. F311, F312). Note that the true dips of structures are known in the core reference frame (e.g., up = +z), but the azimuth of structures is unknown (see the “Methods” chapter). Thus, true dips are plotted on the lower half of rose diagrams for the southeast quadrant only.

Within the ponded lavas, all structures are gently dipping with the exception of high-angle shear veins (Fig. F312). In the massive lavas and sheet flows, structural orientations are variable with steeply dipping fractures and veins and shallowly dipping magmatic structures. Steeply dipping populations of structures dominate the orientations within the transition zone between lavas and upper dikes (Units 1256D-40 to 1256D-43 in Fig. F165). The transition zone is also noteworthy for a possible conjugate set of veins (see Fig. F135 and “Structure orientation”). The sheeted dike complex, in contrast, is dominated by steeply dipping structures with the exception of fractures (Fig. F312). In general, the presence of dikes favors more steeply dipping structures. The distinctive orientation of fractures within the sheeted dike complex suggests that they were initially cooling joints.

During Expedition 312, particular attention was paid to the properties of veins between 1255 and 1367 mbsf, the depth interval over which a change from chlorite- to actinolite-dominated vein assemblages is present. In hand specimen, the veins were described as either dark green (actinolite), light green (chlorite), white (quartz and/or carbonate), or metallic (sulfide). There is no simple relationship where one type or orientation of vein consistently crosscuts another (Fig. F313). From 1255 to 1262 mbsf, shallowly dipping veins cut more steeply dipping ones. From 1271 to 1314 mbsf, more steeply dipping green veins cut more shallowly dipping green veins. Below 1315 mbsf, however, the relationships become much more complex, with a through-going steeply dipping vein set cutting more shallow veins and light and dark green veins crosscutting one another (Fig. F313).

Most veins have widths <1 mm with an exponentially decreasing number of veins with widths >1 mm. In contrast, vein halos follow a roughly gaussian distribution in width with the mode around 2 mm. The difference between vein and halo widths is apparent in a plot of vein width versus halo width wherein each defines different fields (Fig. F314).

Downhole structure of plutonic complex below 1406.1 mbsf

Below 1406.1 mbsf, recovered cores are dominated by gabbroic rocks (Fig. F315). The contact between the granoblastic dikes and the underlying gabbro was recovered in Section 312-1256D-213R-1 (Piece 13). The lithologic contact is sharp, although alteration and recrystallization cover an interval several centimeters wide. There may also be a cataclastic texture overprinted by the recrystallization.

A relatively thick (>6 cm) composite dike of leucocratic gabbro and trondhjemite intruded the gabbro near the top of the body. Below the leucocratic dikes, numerous leucocratic patches are widely distributed, decreasing in size and frequency downward toward homogeneous melanocratic medium-grained gabbros. At the base of the melanocratic gabbro (1421.6 mbsf) are closely packed oikocrysts of clinopyroxene of Unit 1256D-87 (Fig. F315). Thin leucocratic dikelets and patches are abundant below Unit 1256D-87. Shear zones crosscut the leucocratic bands in places. Steeply dipping actinolite veins are also present with accompanying alteration halos.

Below 1458.9 mbsf, dike rocks that contain two pyroxenes were recovered and are referred to as the dike screen (see “Igneous petrology”). A piece from the uppermost part of the dike screen has three orthogonal vein sets: actinolite veins cut by chlorite veins in turn cut by quartz veins. Each vein has planar vein walls. In the lower part of the dike screen, diffuse actinolite veins are present. In contrast, white veins and dikelets in this section have sharp vein walls.

A piece recovered from Core 312-1256D-230R contains a boundary between the dike screen and Gabbro 2 (Fig. F315). In Section 312-1256D-230R-1 (Piece 5), the gabbro intrudes (and crosscuts) the dike screen with flow foliations around the contact. Numerous xenoliths of the dike screen are present in the upper part of the Gabbro 2 unit (see “Igneous petrology”) with their long axes roughly parallel to the flow foliation. Below the xenolith-rich zone, leucocratic dikelets are present along with a few xenoliths of the dike screen. These xenoliths have irregular shapes without significant elongation. Crushed fragments are recovered in three intervals in otherwise almost intact gabbros (Core 312-1256D-231R). Textural banding and flow foliation are also present in the lower part of Gabbro 2.

An interval boundary of the Gabbro 2 unit was recovered in Section 312-1256D-232R-2 (Piece 9). This unoriented whole-round piece has a steeply dipping boundary between coarser grained gabbro and fine-grained gabbronorite of Unit 1256D-94. The fine-grained gabbronorite is intruded by a leucocratic dikelet. Below this unit, fresh basalts of Unit 1256D-95 have a well-defined planar vein system (Core 312-1256D-23R), indicating brittle behavior of the fresh basalt during veining.

Summary

Cores recovered during Expedition 312 from Hole 1256D below 1255 mbsf comprise unstrained sheeted dikes and gabbros. Described structures include fractures, veins, shear veins, breccias, chilled margins and igneous contacts, and alteration and igneous patches. Observed petrofabrics include magmatic flow indicators, local cataclastic domains likely related to intrusive processes and accompanied by alteration and recrystallization textures. Unique to gabbroic rocks are magmatic patches, boundaries between contrasting (now gabbroic) melts and xenoliths, and textural and compositional bands.

Gabbroic and basaltic rocks contain fabrics and structures related to melt transport. Within the gabbros, leucocratic intrusions form bands and patches with local evidence of high-temperature shear. Dikes have chilled margins with flow banding, stretched spherulites, and injections of basalt.

Both gabbros and dikes contain brittle structures that likely formed at high temperatures. In the dikes, veins and breccia domains both cut and are cut by chilled margins. Gabbroic rocks contain domains of intense microfracture of plagioclase, but fractures are largely annealed by secondary plagioclase.

Alteration of gabbros and dikes is pervasive but is enhanced in areas of intrusive contacts. Hydrothermal veins penetrate most of the recovered core, and texture in vein fillings records the fracture opening history under evolving hydrothermal conditions. Veins and vein networks may contain indications for oblique vein opening. Only in sparse and localized shear veins is there evidence for significant noncoaxial strain. Veins are the most common brittle structure in the cores, but joints are also widespread. Most of the shallowly dipping fractures are drilling induced and may have nucleated on preexisting cooling joints and other planes of weakness. These features greatly hindered core recovery.

The population of measured structures systematically changes in orientation downhole. With the exception of joints, all structures become more steeply dipping with depth. At 1406 mbsf, gabbro intrudes the dikes with a 45° dipping contact, below which structures separate into two different populations. Veins below 1406 mbsf are steeply dipping, similar to their orientation in the dikes and lavas. Leucocratic intrusions into the gabbro also have moderate to steep dips. In contrast, igneous contacts, magmatic flow fabrics, and magmatic banding in the gabbros are moderately dipping, similar to the orientation of the contact between the gabbro and the sheeted dike complex.

There is no evidence in recovered cores for significant tilt or strain during accretion and seafloor spreading of Site 1256 crust. Multiple stages of dike and gabbro intrusion, high-temperature (melt-present) shear, and multiple generations of flow at high and moderate temperature can be recognized using structural relationships. The orientations of structures and their mutual relationships show that the strain accommodated by veins, intrusions, and melt migration is intimately related to large-scale magmatic construction of the oceanic crust.

Paleomagnetism

To characterize the paleomagnetic signal and resolve the magnetization components recorded in the igneous rocks of Hole 1256D, we measured and analyzed the magnetic remanence of discrete cubes (~8 and 1 cm3) cut from the working half of the core and selected oriented pieces from the archive half (Tables T43, T44, T45). Remanence data were collected before and after progressive alternating-field or thermal demagnetization. Our primary goal is to assess the roles of different rock types that make up the upper oceanic crust in generating marine magnetic anomalies. Additional goals include aiding structural analysis by providing information on piece orientation.

Most samples have a pronounced drilling overprint, which is characterized by a steep downward direction, and a radial-horizontal component that points toward the center of the core. This was documented during Leg 206 (among others) through measurements of 25 separate pieces of a 1 cm whole round from interval 206-1256D-26R-4, 74–75 cm, located at 443.19 mbsf in Hole 1256D (Wilson, Teagle, Acton, et al., 2003). In the IODP core orientation system, this overprint results in a strong bias in declinations toward 0° for archive-half samples and 180° for working-half samples (see “Core orientation” in the “Methods” chapter). Interpretation of data presented in this chapter was based on the amount of overprint determined by partial demagnetization.

Working-half results

Natural remanent magnetization (NRM) directions of discrete cubes are consistently steep (inclination is at least 45° and usually >60°) and generally have southerly declinations, which is as expected if NRM is dominated by drilling overprint. Most samples tend steadily toward shallower inclinations during progressive alternating-field demagnetization (Figs. F316, F317, F318, F319), although a few reach a stable direction reproduced on several successive demagnetization steps. Directional scatter at high demagnetization fields is fairly minor, in contrast to most of the material recovered from the hole during Leg 206 and Expedition 309.

In order to better constrain the overprint, direction trends during partial demagnetization are represented in an equal area projection for several discrete samples in Figure F320. For each sample, a plane was fit to the successive directions during demagnetization. Intersections of all great circles should be the overprint component shared by all these samples. Drilling overprint is well defined with a southern direction (declination = 179°) and rather steep inclination of 72° (α95 uncertainty = 14.5°). Because of the progressive shallowing of the inclination of most of the samples, we use either the direction at highest demagnetization field or the last direction before the directions begin to scatter at higher demagnetization fields as our estimate of the predrilling magnetization direction (Table T44). This convention differs from that used during Expedition 309, where PCA was applied to a series of 3–8 demagnetization steps before directions scattered (see “Paleomagnetism” in “Expedition 309”). Because PCA is a linear process that preferentially weights vectors of greater magnitude (Kirschvink, 1980), it tends to give directions weighted toward lower demagnetization fields where sample magnetization intensities are higher. In the case of progressive changes on successive demagnetization steps, we prefer to choose the higher demagnetization steps.

Thermal demagnetization was conducted on eight samples with volumes ~1 cm3 (Table T44). Objectives were to help constrain magnetic mineralogy and the blocking temperature of the samples and to check whether thermal demagnetization is a useful supplement to alternating-field demagnetization for removing the drilling overprint. To partially remove the drilling overprint, all samples were alternating-field demagnetized to 10 mT prior to heating, and to remove the effects of laboratory fields between the oven and the magnetometer, the 10 mT demagnetization step was repeated prior to each measurement. Results were fairly consistent between samples for both dikes and gabbros (Fig. F321). Intensity was reduced only slowly with increasing temperature with median destructive temperature between 360° and 480°C. Precipitous intensity drops to nearly complete demagnetization occurred at either the 570° or 580°C steps. This intensity behavior is diagnostic of magnetite with only minimal Ti content. Limited direction change for the 580°–610°C steps suggests that a small fraction of the magnetic carrier may be hematite. Comparison of directions with adjacent samples treated with alternating-field demagnetization shows that thermal demagnetization is less effective at removing the drilling overprint.

Archive-half results

Measurements of entire pieces of the archive half are less directly useful than discrete samples from the working half for two reasons. First, the drilling overprint is stronger and more resistant to demagnetization at the outside of the core, and this outer material is not cut away in the archive half. Second, to preserve some magnetic signal for possible future studies, we did not use demagnetization fields above 40 mT. Nevertheless, archive-half measurements are useful to evaluate the amount of removed overprint in conjunction with working-half samples in order to get cleaner predrilling directions. This is because the radially inward component of the drilling overprint adds differently to the predrilling magnetization for the two halves of the core.

Archive-half NRM directions cluster strongly about a moderately steep and northerly direction. The clustering is substantially reduced by partial demagnetization, with inclinations commonly shallowing to 15°–40° and declinations mostly northwest to northeast. Removal of the overprint is variable downcore. The interval from ~1315 to 1335 mbsf (Sections 312-1256D-184R-1 [Piece 14] through 189R-1 [Piece 4]) has directions still dominated by the drilling overprint after demagnetization, with inclinations steeper than 40° and declinations near 0°. Deeper than 1400 mbsf, much of the overprint is removed, as indicated by inclinations generally shallower than 40° and widely scattered declinations.

Working-half and archive-half integration

In many cases with working-half and archive-half demagnetizations from the same cored piece, the demagnetization direction trends from the two halves appear to converge toward a common direction. Under the assumption that the magnetization in each half is represented by a shared predrilling component and a single overprint component that differs between the halves, it is possible to estimate the predrilling direction by fitting a great circle to each demagnetization trend and determining the intersection of the great circles. Some typical examples of stereoplots and great circle intersections are shown in Figure F322. Inclinations determined by this technique are generally shallower than those estimated from either half separately, and four of the six examples shown cluster between 1° and 7°. Samples from Section 312-1256D-174R-1 (Piece 23) (Fig. F322A) are good examples of how both halves used together could optimize estimation of directions. Although the archive half is strongly overprinted, a trend toward a southwest direction and lower inclination is apparent. The corresponding working half has a demagnetization trend that does not seem to be completed at 50 mT, as no stable direction is reached. Unfortunately, the last demagnetization step performed at 60 mT shows signs of anhysteretic remanent magnetization (ARM) remagnetization, and therefore a proper demagnetization evaluation was difficult. The intersection of the great circles, though, leads to a very shallow inclination (1°) and northwest direction (declination = 346°), which may be closer to the direction prior to drilling. This technique has limited value when the remanent direction is within ~10° of north–south because the planes defined by demagnetization trends are nearly parallel and their intersection becomes very sensitive to noise.

On Figure F323, we plot downcore variations of the best estimate of declination and inclination. The squares represent directions from discrete and archive great circle intersections. Shallow inclination is expected at this paleolatitude, but nearly half of our estimates remain steeper than 20° and none show negative inclination, indicating a remaining drilling overprint. A histogram of inclinations for archive best estimate directions (usually last demagnetization point) and combined discrete and archive best estimate great circle intersections clearly shows the remaining high-inclination overprint. On Figure F324, NRM, ratio of magnetization after 20 mT versus NRM, and susceptibility are represented. The 20 mT/NRM ratio has been shown to correlate with ability to remove the drilling overprint (Wilson, Teagle, Acton, et al., 2003). That correlation continues for Expedition 312 samples, with low values of the 20 mT/NRM ratio for 1320–1340 mbsf correlating with an inability to remove the drilling overprint (Figs. F318, F322D).

Magnetization of a fragment of the borehole wall

In estimating the in situ magnetization of oceanic crust for the purpose of constraining the source of marine magnetic anomalies, it is often useful to have magnetic field measurements from downhole logging. This comparison has been made for Holes 504B and 896A (Worm et al., 1996). One factor that is not known is whether the borehole wall suffers from remagnetization comparable to the cored material.

An unusual opportunity to measure this effect occurred as a side effect of the failure of the drill bit while coring Cores 312-1256D-199R and 200R. The second fishing run, with a milling bit and junk baskets, brought back an obvious fragment of the borehole wall (Fig. F325). The size of the piece is ~10 cm in the downhole and circumferential directions and up to 3 cm in the radial direction. The diameter of curvature of the inner surface of the fragment is ~10 inches, close to the 9.875 inch diameter of the drill bit. The presence of vein fill on some of the outer surfaces dominated by saponite and brown halos adjacent to the veins suggests an original depth between 350 and 820 mbsf, the range of brown halos encountered during Leg 206 (Wilson, Teagle, Acton, et al., 2003) and Expedition 309 (see “Alteration” in “Expedition 309”).

Three approximately radial 1 cm × 1 cm rectangular prisms were cut from this fragment, and each of these prisms was then divided by cutting 0.5–0.7 cm slices parallel to the borehole wall (Fig. F325). Each sample was treated with alternating-field demagnetization at small demagnetization steps, with coordinates defined by treating the inner radial surface of each sample as the cut surface of the working half. Some of the samples show minimal direction change on demagnetization, and others show large direction changes but reach stable directions by 8–20 mT (Fig. F326). NRM intensities are high, ranging 18–47 A/m. Stable directions agree among the samples, suggesting successful removal of overprint. Direction change during demagnetization decreases rapidly away from the borehole wall for Sample 4. It is difficult to differentiate the fraction of the overprint due to normal drilling and the fishing magnet. Interpretation remains equivocal as to whether overprint of the borehole wall is a concern for interpreting the downhole magnetic field. Direction behavior suggests that the overprint is minor, but strong NRM magnetizations permit the possibility of a significant overprint that extends beyond the scale of this fragment.

Summary

Severe drilling overprint hampers attempts to infer the in situ magnetization of upper oceanic crust at Site 1256. Measured NRM intensities are probably higher than the predrilling values, perhaps by a factor of several. Demagnetized inclinations approach the expected nearly horizontal values, but trends continuing to shallow at the last demagnetization suggest that overprint remains. Because of the steep orientation of the overprint, declinations are less sensitive to incomplete removal of the overprint and in most cases should be adequate for approximate strike of structural features. For pieces that can be oriented by core-log integration (see “Digital imaging”), declinations should also be reliable enough to determine magnetic polarity.

Downhole variations in magnetic patterns for Expedition 312 samples are minor, and demagnetization behavior of dikes and gabbros is indistinguishable (Figs. F316, F319). NRM intensity is widely scattered in both dikes and gabbros and the distributions overlap, but intensities in the dikes are generally higher. When viewing results for all of Hole 1256D together, the striking variation is between the zones of low-temperature alteration shallower than 1000 mbsf and high-temperature alteration deeper than 1000 mbsf (see Wilson, Teagle, Acton, et al., 2003) (see “Paleomagnetism” in “Expedition 309”; Fig. F173). The shallow section is highly variable, with a majority of samples severely to completely overprinted, but a small fraction shows minor overprint that appears completely removed by demagnetization. In contrast, the lower section has a consistent, moderate degree of overprint. After maximum demagnetization, a small fraction of the overprint remains. The difference is probably attributed to variations in grain size and oxidation of primary titanomagnetites in the shallow section, contrasted with nearly pure magnetite that is either secondary or altered to exsolve a Ti-bearing phase such as ilmenite in the hydrothermally altered section.

Integration of sample measurements with measurements of the magnetic field in the borehole should allow progress in characterizing crustal magnetization, as should shore-based studies in magnetically cleaner laboratories. The amplitude of marine magnetic anomalies in the area of the site has been satisfactorily modeled by Wilson (1996) with a 500 m thick layer magnetized at 10 A/m. A layer 1250 m thick with a magnetization of 4 A/m would produce an equivalent anomaly. An average predrilling magnetization of 2–5 A/m is within the plausible range for the dikes and gabbros recovered at Site 1256, so they remain candidates for a significant fraction of the source of marine magnetic anomalies.

Physical properties

Measurements of magnetic susceptibility, GRA bulk density, and NGR were made using the MST on whole cores before splitting. Thermal conductivity measurements were made on archive-half cores longer than 7 cm. Minicubes (~9 cm3 in volume) were analyzed for bulk density, porosity, and compressional wave (VP) velocity; a few irregular-shaped pieces were only analyzed for bulk density and porosity. Methods and experimental uncertainties (reported as 1σ) are described in “Physical properties” in the “Methods” chapter.

Because of low recovery in the dikes, the opportunity to obtain discrete samples that could be cut into minicubes in those cores was limited. Most recovered pieces are <4–5 cm long, reducing the reliability of MST results on those pieces. As initiated by the scientists during Expedition 309, in our analysis we consider magnetic susceptibility and GRA values of pieces with lengths >8 cm to reduce the large number of spurious measurements made on gaps between rock pieces.

Minicube density and porosity

Reliable comparison of physical properties measured during cruises that sample rocks from the same borehole depends on careful instrument calibration. To confirm consistency between Expedition 309 and 312 data, we remeasured the wet mass, dry mass, and dry volume of the last six cubes plus one fragment measured during Expedition 309 (Table T46). Values calculated during Expedition 312 for bulk and grain density tend to be lower than Expedition 309 values, but the differences are within instrument precision (<1%) and can be neglected. Differences in porosity values vary; the average absolute difference is 0.2%. Most of these values are <1% (Table T46). Thus, the average difference is ~20% of the calculated porosities. This is large, but it is improbable that a systematic bias of this magnitude exists between porosity values measured during the two expeditions. It is more likely that relative error increases as the low porosity of altered dikes approaches the detection limit of our method. We conclude that interpretations of the variability of porosity values should be treated with caution when values are <1%.

During Expedition 309, bulk density increased with depth in the sheeted dikes as porosity decreased (Fig. F327). The bulk density of Expedition 312 samples (Table T47) continues this pattern (Fig. F328). No noticeable change occurs in values at ~1255 mbsf between Expeditions 309 and 312. For all Expedition 312 samples, average bulk density is 2.97 ± 0.09 g/cm3, grain density is 2.99 ± 0.08 g/cm3, and porosity is 1.2% ± 1.4%.

For Expedition 309 cores, bulk density is inversely related to porosity and grain density is weakly correlated with porosity (Fig. F329C, F329D). Samples from Expedition 312 follow similar trends (Fig. F329A, F329B). However, the two data sets differ in variability. At shallow borehole depths where porosity exceeds 2%, bulk density values at a particular porosity are spread over 0.1 g/cm3, whereas the spread of bulk density values at porosities <1% is only ~0.05 g/cm3. The higher porosity rocks comprise massive and sheet flows recovered above ~1000 mbsf. We suggest that the high variability in bulk density of these units reflects a mixture of alteration mineralogy with a wider range in physical properties than the alteration minerals in the dikes below. This conjecture is consistent with the high variability of grain density in these units (Fig. F329D).

In the Expedition 312 section (Fig. F328), density and porosity abruptly decrease across the contact at 1407 mbsf between granoblastic dikes and the upper gabbro unit. Grain density increases to a maximum value of 3.04 g/cm3 in the lower granoblastic dikes at 1396 mbsf (Section 312-1256D-209R-1; Unit 1256D-80) and then drops to 2.93 g/cm3 in the uppermost gabbros at 1411 mbsf (Core 214R; Unit 1256D-81). Likewise, porosity increases from a low of 0.1% to 1.7% across the boundary. This contrast in physical properties likely reflects changes in mineralogy and grain size (see “Igneous petrology”).

Physical property results also indicate a large degree of inhomogenity in the top section of the upper gabbros (Core 312-1256D-214R; Units 1256D-82 through 84), consistent with the observed variability in petrology and alteration. Both bulk and grain densities vary more in the gabbro than in the lower dikes (Table T47; Fig. F328). Porosity in the top section of the upper gabbros (1407–1425 mbsf) is consistently higher (1%–2%) than porosity in the overlying granoblastic dikes (0%–1%). One sample has a particularly high porosity (8.4%) at 1412 mbsf in an oxide-rich, highly altered zone (Section 312-1256D-214R-1; Unit 1256D-83).

Within the plutonic section, bulk density increases from a low of 2.87 g/cm3 at 1412 mbsf (Section 312-1256D-214R-1; Unit 1256D-83) to a maximum of 3.34 g/cm3 near the top of the Gabbro 2 unit at 1486 mbsf (Section 230R-2). Although this latter value is anomalous, the densities in Gabbro 2 vary between ~3.0 and 3.34 g/cm3 (Table T47) and have larger variance than any other section cored during Expedition 312.

Compressional wave velocity

Very little net change occurs in VP over the ~250 m cored during Expedition 312 (Figs. F330, F331, F332; Tables T48, T49). This observation masks considerable variation occurring in three contiguous intervals with distinctive, asymmetric shape. Each interval comprises an increase at a gradient of ~1 km/s in 50 m and terminates with a step decrease of up to 0.5 km/s in <7 m. The three sharp velocity offsets occur at the top of intrusive units. The shallowest offset coincides with the top of Unit 1256D-76 (between 1324.45 and 1325.28 mbsf in Section 312-1256D-187R-1), where the velocity drops by ~0.4 km/s. The deeper two offsets coincide with contacts between dikes and gabbros. One occurs at the base of the granoblastic dikes (contact between Units 1256D-80 and 81), where the velocity drops ~0.5 km/s between 1404 and 1412 mbsf. The other offset is a drop of 0.3 km/s at the top of Gabbro 2 (Unit 1256D-91). Velocities in the upper sections of the two gabbro units vary considerably, with complementary changes in porosity and density (Fig. F332). Velocity is highest in highly metamorphosed host basalt and lowest and most variable in intrusive gabbros. Thus, intrusive events appear to control vertical changes in velocity.

Compressional wave velocity in Hole 1256D in general, and in Expedition 312 samples in particular, increases with increasing bulk density and decreases with porosity (Fig. F333). This trend is similar to that found elsewhere in oceanic basalts (e.g., Carlson and Herrick, 1990). Velocity variability of samples with porosities <2% is less than variability of samples with greater porosities (Fig. F333). Variability in bulk and grain density is also lower at porosities <2% (Fig. F329). Additionally, the relationship between velocity and porosity appears to improve for samples with <2% porosity (Fig. F333).

Thermal conductivity

Thermal conductivity increases by ~15% through the sheeted dikes in Hole 1256D from ~2.0 W/(m·K) at the top (1060 mbsf) to ~2.3 W/(m·K) at the base (1407 mbsf) (Fig. F331). In Gabbro 1, the median value of thermal conductivity drops to ~2.2 W/(m·K), although variability increases. Thermal conductivity again increases in the Gabbro 2 and lower dike units, reaching 2.5–2.7 W/(m·K) at the base of the hole (Table T50; Fig. F332).

Trends with depth in thermal conductivity appear to correlate with trends in grain and bulk density and porosity (Fig. F331). To test this correlation, we averaged other physical property measurements near the thermal conductivity observations. We used a binning window of 10 m, but the relationships do not differ significantly for windows of either 5 or 20 m. Figure F334 shows reasonably good correlations between thermal conductivity and density and porosity but not with VP . The correlations are best in the sheeted dikes where density is highest and porosity is lowest. The variability of thermal conductivity increases in the upper section of Gabbro 1, with a range from ~1.75 to 2.5 W/(m·K), similar to the increase in variability of velocity and density measured in the minicubes (Figs. F328, F332). In Hole 735B gabbros, large variability in thermal conductivity also occurs, with a range between 2.0 and 2.5 W/(m·K) (Dick, Natland, Miller, et al., 1999).

Magnetic susceptibility

Large gaps occur in observations of magnetic susceptibility because of incomplete recovery. Despite poor coverage, results appear to show that magnetic susceptibility varies downhole with a wavelength of 100–140 m (Fig. F335). For example, magnetic susceptibility increases from ~5,000 SI at 1335 mbsf to 8,000–12,000 SI at ~1370 mbsf and decreases to <3,000 SI at 1425 mbsf. Within this high magnetic susceptibility interval (1335 to ~1370 mbsf), secondary magnetite in thin section increases to 7% ± 2%, from 3% ± 2% for the ~20 m above and below (see “Alteration”). Because magnetite appears in thin sections as a secondary product, it is possible that these cycles represent variations in the nature or degree of alteration.

In the dikes, spikes in magnetic susceptibility reach 14,000 SI (Fig. F336; Table T51), which likely originates from zones of abundant fine-grained magnetic minerals resulting from alteration processes (see “Alteration”). In depth intervals where spikes coincide with high core recovery, the thickness of the high magnetic susceptibility layers appears to be <2–4 m. Higher values (16,000–17,000 SI) occur at the top of the gabbros near 1410 mbsf (Fig. F336). Magnetic susceptibility then decreases to 2,000–4,000 SI in most of the plutonic section. However, values increase to 20,000 SI in a 1.3 m thick unit recovered in Section 312-1256D-230R-1, Unit 1256D-91a (Fig. F337). These values are the highest measured in Hole 1256D.

GRA bulk density

An apparent decrease of 0.16 g/cm3 (~6%) in bulk density as measured by GRA occurs at 1255 mbsf between sampling from Expeditions 309 and 312 (Fig. F338). The cause of this change is uncertain. Neither the bulk density of the minicubes nor the bulk density log (see “Downhole measurements”) change by a similar amount, so it is unlikely that the shift in GRA density reflects a real change in physical properties. However, mineral grain size increases across this boundary (see “Igneous petrology”) and core recovery decreases. It is possible that the decrease in GRA density reflects a subtle change in the nature of dikes indicated by the change in size of mineral grains.

GRA bulk density increases with depth in three segments. The first is from 2.5 g/cm3 in the sheet and massive flows to ~2.8 g/cm3 at 1150 mbsf (Figs. F335, F336). Then a drop of ~0.05 g/cm3 occurs at ~1150 mbsf. From ~1150 to ~1255 mbsf, GRA density increases from ~2.7 to ~2.8 g/cm3. The deepest downhole increase in GRA bulk density (~1255 to ~1503 mbsf; ~2.7 to ~3.0 g/cm3) follows an increase of the maximum peaks in magnetic susceptibility. This higher gradient below ~1255 mbsf may be due to more rapid downhole changes in alteration.

Natural gamma radiation

In most cores examined during Expedition 312, NGR counts are 0–4, which is near background level. Higher values (greater than ~5) were measured in 10 samples (Fig. F336; Table T52). Shipboard chemistry analyses indicate that these samples do not have elevated concentrations of potassium or other elements associated with high radioactivity. Core 312-1256D-189R, however, is the most primitive, and Core 184R is the most altered.

Digital imaging

Rotary coring generally returns azimuthally unoriented samples. Cores can potentially be reoriented by matching features observed in the cores to those imaged by wireline logging of the borehole wall, such as FMS and UBI logs.

Image acquisition

The same protocol to acquire whole-core images was used during Expeditions 309 and 312. Prior to scanning, a red line was marked on each piece to indicate the position of the cut surface of the archive and working halves. Our convention was that the working half of the core was to the right of this line with the core upright. A letter “W” was marked on the working half to minimize the chance of sorting errors after cutting the core. Subsequent observations of the core, such as vein strike or magnetic declination, are oriented by the IODP convention relative to the cut surface (Fig. F15 in the “Methods” chapter). Orienting the red line and structural features (i.e., dip angle and direction of veins and fractures) in the cores relative to a logging image allows orienting core observations to geographic north as measured magnetically by the GPIT on the FMS tool string.

DMT Digital Color CoreScan system (DMT GeoTec, 1996; DMT GmbH, 2000a) was used to scan whole-core piece images. In most measurements, for the purpose of obtaining orientation, all whole-round core pieces that were longer than ~60 mm and that could be rolled smoothly through 360° were imaged. As an exception, fifteen 2 cm wide, “hamburger-shaped” pieces from Core 312-156D-215R were shrink-wrapped and scanned in order to study horizontal fracture using FMS and UBI logs.

For each scanned piece, the length of the piece was measured and the depth to the top of the piece was calculated according to curatorial depth. This information is required at the acquisition stage as it is entered in the DMT software Digicore (DMT GmbH, 2000a). Depths, lengths, and piece numbers of all scanned pieces are provided in Table T53. During Expedition 312, whole-core images of 165 pieces with a total length of 22.8 m were scanned. This represents ~49% of the recovered core.

After whole-round scanning, the core pieces were split and labeled according to IODP convention. The slab image of the archive half of every section was allowed to dry and then scanned, prior to description by the petrologists, using the IODP Geotek Digital Imaging System.

Core image processing

Scanned images were then integrated for each core using the Core Recovery Quality Control program (DMT GmbH, 2000b). Images are plotted on a depth scale according to their IODP curated depths, leaving appropriate gaps where material was not scanned or not recovered. The depth profiles are output as EMF files, which can then be opened, edited, and saved in Adobe Illustrator format.

Archive-half slab images were added to the Adobe Illustrator files and aligned in curatorial depth alongside the whole-round images for comparison between the external and internal features of the core (see “Supplementary material”). These files are intended for use as a template for detailed postcruise structural and core-log integration studies, in particular for whole-round core and downhole log image correlation. They will aid in determination of core depth with respect to downhole logs and core reorientation with respect to geographic north as measured magnetically by the GPIT on the FMS tool string.

Core-log integration

Only a few preliminary attempts at matching core images to logging images have been made because of limited time between the collection of downhole logging data and the end of the cruise. Some of these attempts, using the methods described in Haggas et al. (2001), show potentially good matches between unrolled core images and the FMS and UBI data from Hole 1256D. Figure F339 shows examples of a whole-core image, an archive-half slab image, an FMS log, and a UBI log. The examples were produced during Expedition 312 using the largest piece of a high-recovery core from Expedition 309 (Section 309-1256D-85R-1). Although there exists an offset in the depth between core observations and logs, the spacing and dip of the fractures can be collected, corrected, and correlated convincingly between core images and FMS images, with the cutting line on the north side of the core. Juxtaposition of shallowly dipping veins (0, 1, 3, 4, 5, and 6) and a distinctive sinusoid curve (2) are fairly matched to those features observed on the FMS and UBI images. Changes in resistivity/​conductivity corresponding to the degree of fracturing in the pieces may support the depth matching (green ellipses). The depth shift between the whole-core image and the FMS and UBI images is ~1 m, presumably due to incomplete compensation of heave and/or tide.

Downhole measurements

Downhole measurements in Hole 1256D were conducted after the conclusion of coring operations at Site 1256 during Expedition 312. Wireline logging operations during Expedition 312 built on the success of Leg 206 and Expedition 309 and provide for the first time in situ physical property measurements of the sheeted dike–gabbro transition.

Primary logging objectives included refining the igneous lithostratigraphy and examining variations in seawater-basalt alteration as a function of depth at a superfast spreading center. Logging data also allow direct correlation of wireline measurements with discrete laboratory measurements on recovered core. Furthermore, core recovery during drilling igneous basement is often incomplete and biased, with weaker lithologies preferentially lost. Wireline logging provides continuous data over most intervals, including those with low recovery, and consequently complements coring.

Preliminary log interpretation of Leg 206 and Expedition 309 logging data

Wireline operations during Leg 206 and Expedition 309 (Wilson, Teagle, Acton, et al., 2003) provide in situ physical properties of the upper part of the oceanic crust at Site 1256 to a depth of ~1220 mbsf. Well logging during Expedition 312 was a continuation of these efforts and provides further constraints on the physical properties of deeper sections of the oceanic crust, in particular the sheeted dikes–gabbro transition.

At the beginning of Expedition 312, a tentative reconstruction of the lithology at Site 1256 was carried out using well-logging data acquired during Leg 206 and Expedition 309. The purpose of reconstructing lithology from well-logging data is to translate physical measurements from logs into lithologic terms. Each type of rock is characterized by a set of log responses, which distinguishes it from other types of rocks. These log responses are dependent, among other parameters, on the composition, texture, and mineralogy of the rocks. For this purpose, individual well-logging data points were averaged to 10 m running averages, and the results are illustrated in Figure F340. The objective for applying this reduction was to determine the principal relations between wireline logs and core lithology and to improve the evaluation of possible trends or significant lithologic differences.

Three principal rock types were distinguished during Leg 206 and Expedition 309, with sheet flows or brecciated basalts the most common, followed by massive units and then pillow basalts. Pillow basalts were described only from the upper borehole between ~365 and 375 mbsf. Relative differences in the log response between these structurally different rocks are demonstrated in eight different sections of Hole 1256D (Fig. F340; Table T54). These sections represent areas with higher sample recovery and/or good borehole conditions indicated by caliper readings <12 inches. It is evident that highly fractured lithologies like pillow and brecciated basalts display higher natural radioactivity compared to massive units. These fractured units are also characterized by variable porosities and densities with values well above 5% and below 2.9 g/cm3, respectively. Compressional velocities for these units vary from 3.2 to 5.5 km/s. Pillow basalts may be distinguished from brecciated lithologies by resistivities ≤10 Ωm. However, a clear discrimination between these units using well-logging data alone remains uncertain. Examples for massive units are highlighted in the depth intervals 316–338, 472–490, 819–833, and 1120–1140 mbsf (Fig. F340). These units are clearly distinguished from the previously described lithologies by high compressional velocities (>5.5 km/s) and densities (~2.7 g/cm3) and increased resistivity (usually >100 Ωm) and correlate with low porosity (<12%) and NGR emissions (<4 gAPI).

The most compelling change in log response is observed below the transition zone (~1060 mbsf) in the sheeted dikes. Natural radiation in these rocks remains relatively constant with values generally <3 gAPI. This low value may reflect a lack of K-bearing minerals (e.g., saponite), which are the main sources of naturally occurring radioactivity in these rocks (see “Alteration” and “Geochemistry”). Increased bulk density, compressional velocity, and electrical resistivity demonstrate a clear change in lithology and show the highest values obtained in Hole 1256D. Furthermore, resistivity data recorded with the DLL tool (see the “Methods” chapter) demonstrate a strong decoupling between shallow (LLS) and deep (LLD) resistivity below 1080 mbsf. LLS measurements have the same vertical resolution as the LLD but respond more strongly to that region around the borehole affected by drill water invasion. Caliper readings from 1080 to 1211 mbsf are on average 10.98 inches (±0.5 inch), indicating good borehole conditions, and shallow resistivity measurements are consequently less influenced by fluid invasion. It is therefore unlikely that fluid invasion is solely responsible for the observed decoupling of both resistivity measurements. Similar differences between shallow and deep resistivity in Hole 504B have been attributed to an anisotropic distribution of pore space in the rock (Pezard and Anderson, 1989). In the case of a subvertical network of conductive fractures, shallow resistivity is more strongly reduced than deep resistivity. It is very likely that resistivity data obtained in Hole 1256D also indicate a dominance of vertical features in the sheeted dikes.

Operations

Following completion of coring operations in Hole 1256D at a depth of 1507.1 mbsf, the hole was conditioned by displacing a mixture of seawater and 210 bbl of 8.9 ppg sepiolite. A wiper trip was run, and a logging BHA with logging bit was run into the hole. The base of the BHA was set at 280 mbsf, and a total of six tool strings were deployed (Fig. F341).

Triple combo tool string

The first deployment consisted of the triple combo tool string, which contained the HNGS, APS, HLDS, DLL, and TAP tool.

After the tool string was lowered ~2500 m into the pipe, the wireline winch started to stall. Inspection of the chain drive indicated that increased tension may have forced a slight displacement of the cable drum. This required adjustment of the cable drum to prevent serious damage to the winch, and after 2 h deployment of the tool string continued. Another problem during this deployment was that the connection to the DLL tool could not be maintained. Several attempts failed to reestablish connection to this tool, so it was decided to complete the logging run and test the tool at a later stage and to possibly rerun this tool at the end of the logging operations. At a depth of 1440 mbsf, the cable head tension decreased, indicating that the tool string reached its maximum depth, ~67 m above the total cored depth. The first uphole logging run was conducted from this depth and reached 343 mbsf. A repeat pass was performed from 1438 to 1080 mbsf and was also successfully completed. The recorded NGR during the repeat pass was slightly higher than the main pass. Although nearly 5 h elapsed between the beginning of the main pass and the repeat pass, the formation was possibly still affected by the neutron activation of the APS tool.

Vertical seismic profile

The second wireline logging operation was the acquisition of a VSP in Hole 1256D. The VSP was planned to provide seismic interval velocities that can be correlated with the drilled rock sequences to place the borehole in context of the seismic refraction velocity structure of the oceanic crust. These data can assist in defining the seismic boundary between seismic Layers 2 and 3, although this boundary lies beneath the current bottom of Hole 1256D.

Experimental set-up

The air gun depths and generator-injection (GI) chamber configuration for the seismic experiment were chosen after evaluating the Leg 206 check shot data. It appeared that the amplitudes of the first arrivals decrease with increasing depth of the seismometer in the borehole. This reflects attenuation in the upper crust and may be responsible, along with an increase in noise, for the observed low signal-to-noise ratio acquired from deeper stations. The source for the noise remains unclear but appears to be random because stacked traces from these seismometer depths have reasonable signal-to-noise ratios and a clear first arrival. Stacking the seismic traces eliminates enough noise that arrivals can be detected. The air gun used during Leg 206 was configured in a true GI mode, having both a 75 inch3 generator and injector chamber. The Expedition 312 VSP survey was planned to reach levels ~690 m below the deepest station reached during Leg 206. Modification of the GI gun was consequently necessary to increase the first arrival signal in sections of the borehole below 750 mbsf. The GI chamber configuration was changed into “harmonic mode” at 150 and 105 inch3, respectively. Furthermore, the air gun was suspended 7 mbsl, which resulted in a shift of the peak frequency from 80–120 Hz to ~50 Hz. This energy shift to lower frequencies also allowed penetration to deeper levels of the oceanic basement.

The air gun and buoy were deployed by the port crane of the JOIDES Resolution, and the gun was tethered to the ship forward and aft of the crane to prevent rotation. The air gun was operated at 2000 psi, and a blast phone for detecting the shot instant was rigged 2 m below the air gun. The actual gun-to-hydrophone distance may have varied during the experiment owing to currents. However, the source signature, which includes the reflection of the pressure field from the sea surface, remained relatively constant.

Experimental narrative

Rigging up the VSI was completed on 20 December 2005. The VSI tool was lowered to the end of the pipe and halted in a safe position. However, the tool was not lowered into the open hole until daylight (0541 h) on 21 December. TD of 1433 mbsf was reached at 0733 h, and the check shot survey started after 1 h in accordance with the marine mammal protocol. Weather was calm and clear and the sea state was calm, providing excellent conditions for the mammal watch and the survey. The survey began with a 0.5 h ramp-up of the air gun, beginning at 500 psi and gradually increasing the pressure to 2000 psi. No mammals were sighted during the entire survey, and the survey continued without interruption. The first clamping was attempted at 1433 mbsf, but because of the absence of a first arrival signal, the second clamping was placed 50 m higher at 1383 mbsf. After receiving an excellent first arrival signal, the remaining clampings were attempted at 58 depths, 22 m apart. These clampings cover the entire basement section of Hole 1256D. At each station, 11–25 shots were taken and selected shots were added to the stack. All shots from a clamping station were recorded in one file and the stacks from each clamping were stored in separate files. Three of the deepest Leg 206 clampings (610, 700, and 729 mbsf) were reoccupied to provide data overlap and comparison with the previously performed check shot survey.

Formation MicroScanner-sonic tool string

The third tool string deployment during Expedition 312 was the FMS-sonic, which consisted of the DSI, SGT, GPIT, and FMS tools. Rigging up started at 1930 h on 21 December 2005 and was completed at 2030 h. The tool string was run into the hole after successfully concluding a test of the FMS caliper arm function on deck. TD of 1437 mbsf was reached, and the main pass started from that depth. Caliper arms of the FMS tool were opened at 1431 mbsf, and data were received from all tools. After 15 min, the main voltage dropped and the main current increased to 1600 mA. The data channels of the FMS and GPIT tools returned random values, and the only functioning tool was the DSI. As it was not possible to regain control over the FMS tool, the decision was made to stop the logging run at 1309 mbsf. To avoid damage to tool and borehole, the FMS was manually bypassed to close the caliper arms and troubleshoot the tool. The only way to test if the caliper arms were closed at this stage was to monitor the head tension on the screen when the tool string was carefully moved up and down again. The head tension remained the same in both directions, indicating that the caliper arms closed successfully. However, as it was not possible to gain control over the FMS tool and the main current continued to be extremely high, the logging run was aborted and the tool string returned to the deck to avoid damage to the instruments. The FMS tool string successfully entered the pipe and rig-down was completed 3 h later.

Ultrasonic Borehole Imager and Dipole Shear Imager tools

A fourth tool string consisting of the UBI, GPIT, and SGT tools was run next. The DSI, usually part of the FMS-sonic tool string, was added because obtaining high-quality velocity data was a high priority in the logging operations. For the first logging run, the DSI tool was configured using the standard frequency R15 to determine the formation resistivity. The tool string reached a TD of 1430 mbsf, and the first pass logged to 1099 mbsf. After completion, the tool string was lowered for a repeat pass covering the 1432–1322 mbsf interval using the same standard frequency for the DSI. A second main pass was run using an R3 firing rate with a medium frequency, and both dipoles of the DSI were set in the cross receivers mode. The second main pass was run from 1433 to 1089 mbsf, and after completion the tool was returned to deck and rigged down.

Formation MicroScanner-sonic tool string, second run

After successfully troubleshooting the FMS, the tool was combined with the SGT, rigged up, and lowered into the hole at 1558 h on 22 December 2005. The TD of 1437 mbsf was reached at 1949 h, and two main passes were successfully run, covering the intervals 1437–1089 and 1436–1101 mbsf. The FMS caliper arms encountered an obstruction at 1354 mbsf on the first pass and temporarily closed but reopened after 5 m. No further problems were met during this run. The obstruction was not encountered during the second FMS pass, and the tool recorded high-quality data during both runs. After completion of both passes, the tool was recovered and rigdown completed at 0340 h on 23 December 2005.

Temperature/​Acceleration/​Pressure and Dual Laterolog tool string

The last tool string during Expedition 312 included the TAP, DLL, and SGT tools. The tool string was slowly run into the open borehole at 900 m/h to provide downhole and uphole recording of data. Furthermore, temperatures recorded at the beginning and end of the logging operations provide excellent ways to estimate the equilibrium temperature at the bottom of the hole. TD of 1440 mbsf was reached at 0829 h, and the entire borehole was logged up to 290 mbsf. The tool string entered the pipe without problem at 1000 h, and the entire rigdown was completed at 1212 h. Logging operations in Hole 1256D terminated with rigging down the wireline and clearing the derrick area of tools at 1300 h on 23 December 2005.

Processing and data quality

Following acquisition, logging data were transmitted to LDEO for depth and environmental correction processing (see “Downhole measurements” in the “Methods” chapter), and the processed data were then returned to the ship. Depth matching was done by matching all logging passes to the first triple combo pass.

The principal logging results are shown in Figure F342. Borehole conditions were good during the six logging runs, and no major washouts or obstructions were encountered. However, the total cored depth of 1507.1 mbsf could not be reached because of ~67 m of fill after the last pipe trip. Caliper readings from the triple combo and the FMS indicate excellent borehole conditions in the newly drilled section below 1255 mbsf, with a diameter typically between 10 and 11 inches. However, when compared to Expedition 309 caliper data from logging phases 1 and 2, the progressive enlargement in the upper borehole sections due to continued drilling is apparent (Fig. F342). In the intervals 530–600, 650–700, 920–960, and 1050–1065 mbsf, the hole caliper increased by as much as 10 inches. The maximum extension of the caliper of 22 inches was reached between 920 and 960 mbsf, indicated by a constant reading in this depth range. However, the average hole diameter measurements below 1255 mbsf are 10.57 inches for C1 and 10.31 inches for C2. This demonstrates nearly radial borehole condition throughout the new section. The only section where larger borehole diameters up to 14 inches were encountered was at 1300 mbsf. The FMS caliper C1 also measured over 12 inches at 1322 mbsf and from 1410 mbsf to TD, indicating uneven borehole dimensions. An enlarged borehole affects measurements that require eccentralization and/or good contact between the tool and the borehole wall, in particular the APS, HLDS, UBI, and FMS tools. As the enlargements of the newly drilled section are minimal, the aforementioned measurements should not have been strongly affected.

All tools provided high-quality data and demonstrate an excellent overlap with logging runs from Expedition 309. Slight differences in absolute values of gamma ray radiation measurements of Expeditions 309 and 312 are related to the statistically based nature of the measurement. Differences observed in density and porosity between expeditions are related to the change in borehole diameter after coring between individual logging runs. Hole deviation measured during Expedition 309 at 1200 mbsf reached 4.3°, and the hole azimuth varied between 250° and 290°. During Expedition 312, hole deviation increased by 1.1° and reached a maximum of 5.4° at 1427 mbsf. The hole azimuth ranged between 254° and 281° but was, on average, 271.5°.

Preliminary processing of the acquired FMS and UBI data provided high-quality data output. The slow logging speed (120 m/h) used for the UBI main pass during Expedition 309 resulted in an overlap of recorded data and led to relatively poor images of the borehole wall. An increase in logging speed for the UBI repeat pass to 250 m/h provided higher quality borehole images. The same approach was applied to all UBI logging runs during Expedition 312 and supplied high-quality data. The two FMS passes were depth-matched to the first triple combo run of Expedition 312. Although the FMS images can be corrected with confidence, the UBI images show artifacts of sticking. Acoustic and resistivity images were statically and dynamically normalized during conversion into color images. Coverage of the borehole wall by the two FMS passes is good and provides a substantial supplement to the UBI images.

Results

Naturally occurring radioactivity was measured continuously on each logging run and was used to depth match each logging run relative to the triple combo logging. Overall, total gamma ray remains relatively constant and well below 4 gAPI in the section logged during Expedition 312 (Fig. F343). A minor increase above 4 gAPI occurs at ~1210 mbsf, and values reached 6 gAPI between 1300 and 1330 mbsf. The increase at 1210 mbsf is associated with a washout where the caliper recorded ~13 inches, whereas the increase below 1300 mbsf corresponds with elevated electrical resistivity.

The DLL tool recorded the shallow and deep electrical resistivity of the formation to a TD of 1416 mbsf. Generally, resistivity increases with increasing depth, but this trend is interrupted at several depth intervals. Below 1250 mbsf, shallow and deep resistivity measurements are between 500 and 10,000 Ωm and 1,000 and 40,000 Ωm, respectively. These readings are higher than values reported from Expedition 309 and may indicate a change in lithology and/or state of alteration. However, formation deep resistivity measurements may be considerably higher than 40,000 Ωm, as this is the maximum reading for the DLL tool. Strong decoupling between shallow and deep resistivity measurements described from the top of the sheeted dikes continues from 1200 mbsf to TD. Decoupling between the two resistivity measurements increases below 1325 mbsf. Although both measurements are affected, deep resistivity displays a much stronger increase to a depth of 1342 mbsf. Shallow resistivity increases gradually to the TD and remains just above 10,000 Ωm. As the LLD tool was operating close to its maximum range (40,000 Ωm), the difference in resistivity between the LLS and LLD curves at high resistivity may be exaggerated. Another small increase in both measurements can be identified at 1370 mbsf. As mentioned above, shallow and deep resistivity measurements have the same vertical resolution but respond differently to the close vicinity of the borehole wall. Both resistivity measurements decrease by an order of magnitude at ~1300 mbsf, which corresponds to an enlarged borehole. Decreases unrelated to changes in borehole size can be observed at 1220, 1240 1265, 1270, 1290, and 1320 mbsf. At these depths, resistivity is affected by brecciated lithologies or intrusive contacts. These changes in electrical resistivity correlate with increases in neutron porosity and decreases in density. The general increase of resistivity with depth may be related to the change in metamorphic grade (see “Alteration”). Based on variability and magnitude of the electrical resistivity of the sheeted dike complex, the lithology may be divided into four sections: 1060–1155, 1155–1275, 1275–1350, and 1350–1407 mbsf (Fig. F342).

Although overall values of density and neutron porosity range from 1.5 to 3.1 g/cm3 and 2% to 75%, respectively, most variation remains small in the newly cored section of Hole 1256D. Neutron porosity is mostly below 7%, reaching an unrealistically high value of 75% at 1300 mbsf. Density values are usually close to 2.9 g/cm3, but a small increase in density occurs below 1250 mbsf, paralleled by a decrease in neutron porosity. Distinct anomalies correlate with enlarged borehole diameters related to washouts. Average densities of the sheeted dike complex and the granoblastic dikes are 2.89 and 2.99 g/cm3, respectively. Below the granoblastic dikes, density drops to an average of 2.95 g/cm3 in Gabbro 1. This change in density is accompanied by a decrease in compressional velocity from 6.2 to 4.6 km/s. Density and compressional velocity measured on discrete samples also drop at these two depths (see “Physical properties”).

The VSP was shot in Hole 1256D to determine interval velocities and record seismograms for further analysis of seismic properties of the upper ocean crust. Clear compressional wave arrivals were obtained at 1383 mbsf, demonstrating that the GI air gun in the 150/105 harmonic configuration provided sufficient signal to penetrate to midcrustal depths. Data from the three orthogonal geophones in the VSI tool air gun were recorded with a 1.5 s delay at 1 ms sampling interval. The signal from the blast phone suspended 2 m below the gun was recorded without delay as a fourth channel. We attempted to collect data at stations spaced 22 m apart in the open borehole. The most serious problem was poor anchoring to the borehole wall, indicated by harmonic ringing in the seismograms. In many cases, we were able to move the VSI tool 1–5 m uphole and obtain secure clamping, but stations at 1054, 934, 811, 587, and 444 mbsf had to be abandoned, leaving gaps up to 45 m between stations (Table T55).

At each clamping station, shots were viewed and a decision made whether to add the raw seismogram to the preliminary stacked seismogram for that station. An automatic picking algorithm provided traveltimes, but picking did not always select the true compressional arrival. Both inflexibility in the stacking process and lack of control on the picking routine contributed to traveltime errors. Because the relative change in depth between stations is probably accurate to <0.5 m (~2%), traveltime errors make the largest contribution to errors in the preliminary interval velocities listed in Table T55. These errors can be quickly detected because an anomalously short traveltime at one station must produce a high velocity on one side and a low velocity on the other. For example, the low velocity of 5.1 km/s compared to logging and minicube velocities at 1207–1229 mbsf is balanced by a high velocity of 6.3 km/s at 1185–1207 mbsf. Similarly, the low/high velocities in the Leg 206 interval velocities at 610–660 and 570–610 mbsf (red line in Fig. F344) are likely to be a traveltime error at the 610 mbsf station. The pair of low/high velocities from Expedition 312 VSP stations at 647–701 mbsf appears to be due to a traveltime error, but the swing in velocity is matched by a change in velocity from the Expedition 309 sonic log.

Two unusually high interval velocities are not matched by low velocities at neighboring stations. We obtained a velocity of 7.6 km/s between 1339 and 1361 mbsf and a velocity of 6.5 km/s at 880–903 mbsf. Borehole data indicate the deeper unit is a low-recovery interval of aphyric microcrystalline basalt (lithologic Units 1256D-77 and 78; Cores 312-1256D-190R through 195R), whereas the shallower unit is an aphyric microcrystalline sheet flow (lithologic Unit 1256D-33a; Cores 309-1256D-94R through 98R). Because the velocities of these lithologies are unlikely to be >6 km/s, the only plausible geologic explanation is the presence of a nearby high-velocity intrusive body that was not sampled by drilling. Alternatively, it is possible that traveltime errors at adjacent stations produced these velocities.

In general, VSP interval velocities parallel trends in the sonic log and the shipboard velocity measurements on recovered rock samples (Fig. F344). Although the velocity magnitude differs among the various methodologies because of different frequencies of sound and different confining pressures, the trends with depth are similar. This similarity demonstrates the fundamental dependence of velocity fluctuations in uppermost crust on the primary eruptive process and the increase in velocity with depth in Layer 2 on the increasing density of the rocks due to progressively higher temperature alteration and metamorphism.

Preliminary analysis of resistivity and sonic image data indicates high-quality data were obtained over the logged borehole section. The quality of the data is apparent in Figure F345, covering the dike/​plutonic section boundary between 1402 and 1419 mbsf. Directly above the boundary, the formations are characterized by randomly oriented fractures, whereas fractures in the gabbroic section are regular. However, other data suggest that the vertically oriented high-amplitude zones in the UBI image may be tool-related artifacts rather than characteristics of the formation, as these features are not observed in FMS images. FMS and UBI images are complementary and are crucial in reorienting cores. Oriented images can provide essential information on reconstructing the relative orientation of fractures and veins. Features presented in the UBI image at 1410 and 1418 mbsf have a northeast-oriented plunge and an approximate dip between 35° and 40° that may represent fractures. The same features are also evident on the resistivity image, where they represent zones of high conductivity. The anomaly at 1410 mbsf could possibly be the dike/​gabbro boundary, but this identification requires postcruise confirmation.

The TAP tool was deployed at the beginning and end of the logging operation to gain information on the thermal rebound in the borehole after coring ceased. During the first logging operation, the TAP tool was lowered to TD twice (Figs. F341, F346). It appeared that during this deployment the bottom hole temperature increased from 64.24° to 67.90°C within the time frame of 5 h, 10 min. Temperature perturbations are visible between 900 and 950 mbsf and 1350 and 1400 mbsf with negative deviation from the temperature profile. These negative temperature anomalies indicate a slower return to equilibrium temperatures and may be due to greater seawater invasion during the drilling process. This interpretation is supported by enlarged borehole diameters between 900 and 950 mbsf. However, not all borehole sections with increased diameter (e.g., 810–840 mbsf) display strong negative temperature deviations. In particular, in the upper borehole, this may indicate such sections experience a more rapid thermal rebound than the deeper borehole sections. The second TAP tool deployment including the Environment Measurement Sonde (EMS) temperature tool recorded 86.6°C at the bottom of the borehole 68 h, 29 min after the previous deployment (Fig. F346). This demonstrates an increase of 17.6°C within a relatively short period of time. The EMS temperature tool did not measure temperatures between 660 and 820 mbsf during the downhole measurement but parallels the temperature profile recorded by the TAP tool for the remaining borehole.