PHYSICAL VOLCANOLOGY AND IGNEOUS PETROLOGY

We encountered basement at 57 mbsf at Site 1206 (Hole 1206A) on Koko Seamount and drilled 278 m into a sequence of lava flows, hyaloclastites, volcaniclastic sandstone, and limestone, of which 144 m (52%) was recovered. Although the basement of Koko Seamount has never been drilled prior to Leg 197, it was dredged during the Thomas Washington expedition Aries 7 (Davies et al., 1972; Clague and Dalrymple, 1973) and tholeiitic basalt, alkalic basalt, hawaiite, mugearite, trachyte, and phonolite were recovered.

A biostratigraphic age of early to middle Eocene (nannofossil Zones NP14 and NP15) was assigned to sediment immediately overlying basement at Site 1206, giving a minimum age of 43.5-49.5 Ma (see "Biostratigraphy"), which is consistent with the 48.1-Ma K-Ar radiometric age for samples of dredged alkalic lava from Koko Seamount (Clague and Dalrymple, 1973). Basement rocks at Site 1206 consist of 15 lava flow units, 4 volcaniclastic sandstone horizons, and 3 thin limestone interbeds. The basaltic eruptive units often include hyaloclastite breccia or multiple lava flow lobes alternating with hyaloclastite. The basalt is mainly tholeiitic, although two samples are alkalic in composition.

Macroscopic Description

This section describes the different lithologies of the basement sequence in Hole 1206A. Detailed descriptions of each core section can be found in "Site 1206 Core Descriptions".

The basement in Hole 1206A was divided into 22 lithologic units (Tables T2, T3; Fig. F2) based on major changes in lithology such as the occurrence of weathered flow tops, sedimentary interbeds, or basal flow breccia, or abrupt changes in vesicularity. The basement rocks are mainly lava flows (Units 1, 2b, 4, 5, 6, 7, 10, 13, 17, 18, and 21 and parts of 8, 11, and 14) with associated hyaloclastite breccia beds (Subunits 2a and 2c and Unit 19 and parts of Units 8, 11, and 14) and a few interbeds of limestone (Units 3, 9, and 15) or sandstone (Units 12, 16, 20, and 22). The lava flows are dominantly olivine-phyric to aphyric basalt, although a few units also contain up to 3 modal% plagioclase phenocrysts. Olivine phenocrysts usually make up <5 modal% of the porphyritic basalt, except in Unit 17, and of olivine accumulation layers in Units 1, 4, and 6, which contain up to 25 modal% olivine phenocrysts. Plagioclase and olivine and, occasionally, clinopyroxene may be present together as glomerocrysts. Groundmass minerals include plagioclase, clinopyroxene, black oxides, minor glass, and, rarely, olivine.

Some lava flows are lobed, but most others consist of a single lobe (i.e., cooling unit) that includes a vesicular flow top, a massive interior, and a vesicular flow base often featuring pipe vesicles. Many of the flows are highly vesicular (up to 60%), with 1- to 10-mm-diameter vesicles, especially close to unit boundaries. Vesicles are sometimes filled with secondary minerals, including clay, carbonate, and zeolite. The basalt is moderately to highly altered near the top of the section, but only slightly to moderately altered deeper in the section (see "Alteration and Weathering"). Olivine phenocrysts are the first phase to show signs of alteration and are often completely replaced by serpentine, iddingsite, celadonite, or talc. However, in several units, much of the olivine remains unaltered. Unaltered glass persists in some of the lobe margins and hyaloclastite units.

The hyaloclastite beds are most commonly present at the base of lava flow units, except for Unit 2, where it also caps the lava, but hyaloclastite layers are also interbedded with lobes of aphyric basalt in Units 8, 11, and 14. They are clast-supported hyaloclastite lapilli breccia consisting of angular glass and basalt lapilli to breccia fragments in a matrix of fine lapilli. Large basalt clasts are also present, and in some cores it is difficult to determine whether these are large clasts or discrete flow lobes in the breccia sequence.

Some thin intercalations of sediment (limestone, volcaniclastic sandstone, and grainstone and a red-brown deeply weathered flow top) are also present. These are summarized in Table T2. Detailed descriptions of these intervals are given in "Site 1206 Core Descriptions", but some general comments can be made. In the upper part of the basement sequence (Cores 197-1206A-2R to 27R) the sediment interbeds consist of layers of matrix-supported calcilutite, calcarenite, and limestone with bioclasts and miliolid foraminifers. In the lower part of the basement section, rather thicker layers are grain-supported calcareous grainstone, volcaniclastic grainstone, sandstone, and conglomerate. Bioclasts include cylindrical particles, miliolid foraminifers, and millimeter-sized fragments of coral, sponge spicules, and bivalves. Moderate bioturbation is limited to vertical burrows at infrequent intervals. The size range, sorting, and biogenic components of these sediment interbeds suggest deposition in a nearshore environment.

Physical Volcanology

Volcanic Succession

The basement drilled at Site 1206 is composed of a 278-m-thick succession of 22 lithologic units (Fig. F2; Table T3). We have identified 15 units as lava flow formations, and 7 units are sedimentary in origin.

The sedimentary units consist of fossiliferous vitric-lithic sandstone and gravelstone along with minor limestone intervals and exhibit textures, structures, and fossil assemblages indicative of deposition in a wave-dominated nearshore environment (see "Site 1206 Core Descriptions"; Table T2).

The volcanic rocks make up ~88% of the basement section recovered at Site 1206 (Table T3). They are broadly similar in composition to the magmas erupted during the shield stage of Hawaiian volcanoes (see "Geochemistry"). We divided the volcanic succession into 15 lava flow units (as defined in Table T3 in the "Explanatory Notes" chapter) and 10 eruption units (see Tables T3, T4). However, subsequent paleomagnetic and geochemical measurements have raised some concerns about the validity of this division in intervals of poor recovery (i.e., Unit 4: Cores 197-1206A-8R to 14R and Unit 6: Cores 17R to 19R) (Table T5). These uncertainties are addressed in detail below.

The top part of Unit 4 (i.e., Sections 197-1206A-7R-2 to 7R-4) consists of four 0.1- to 1.1-m-thick pahoehoe lobes containing olivine phenocrysts (5-7 modal%) and plagioclase microphenocrysts (2-3 modal%). Sections 197-1206A-8R-1 and 8R-2 of this unit contain broken core consisting of highly and sparsely vesicular lava fragments. No distinct flow lobe structures are discernible. Furthermore, in Section 197-1206A-8R-1, Pieces 21, 22, 24, 25, and 27 consist of fractured lava with detrital carbonate filling the cracks, whereas Pieces 23, 26, 29, and 30 consist of laminated to massive calcareous lithic-vitric sandstone. The lava in Sections 197-1206A-7R-2 to 8R-1 contains olivine phenocrysts (1 mm) along with plagioclase microphenocrysts. Plagioclase microphenocrysts are not present in the lava from Section 197-1206A-8R-2 (below Piece 9) and in subsequent cores in Unit 4. However, this change in plagioclase microphenocryst content of the lava does not coincide with the presence of detrital carbonate as fracture fill. It is present across an inferred lobe boundary placed at 36 cm in Section 197-1206A-8R-2 on the basis of changes in vesicle fabric and lava crystallinity.

Because of drilling problems (i.e., a clogged bit) the recovery in Cores 197-1206A-9R through 13R was exceptionally low (recovery = 0.43%-22%) (Table T5). The lava in Core 197-1206A-14R represents the material recovered from the clogged drill bit, and therefore its stratigraphic significance is ambiguous. Consequently, the internal architecture of the lava in Cores 197-1206A-9R through 14R is unknown, although the vesicularity and vesicle fabric of the lava in these cores is consistent with a pahoehoe flow origin. The olivine phenocryst size (3-5 mm) and abundance (7-25 modal%) of the lava in Cores 197-1206A-9R to 14R is significantly greater than those found in the upper part of Unit 4 (Cores 197-1206A-7R to 8R). Thus, it is possible that this unit is made up of lava from more than one eruption, and this uncertainty will be addressed by shore-based studies.

Unit 6 consists of multiple 0.2- to 2.0-m-thick pahoehoe lobes, characterized by smooth glassy lobe margins, and the overall architecture of the unit is consistent with that of a compound pahoehoe flow (Table T4). A significant change in size and abundance of olivine phenocrysts occurs across Cores 197-1206A-17R to 18R: from <1 mm and 5-8 modal%, respectively, in Cores 16R to 17R to 4-5 mm and 10-20 modal%, respectively, in Core 18R to Section 19R-1 (see "Site 1206 Core Descriptions"). This change is coupled with significant deficiencies in recovery (Table T5) and changes in magnetic inclination (see "Paleomagnetism and Rock Magnetism"). This latter evidence suggests that this lava unit may consist of two separate eruption units (i.e., flow fields).

Of the 15 lava flow units, 11 (66%) are composed of subaerial basaltic lavas and 4 are composed of basalt hyaloclastite breccia with lava lobe intervals (34%) (Tables T3, T4, T6). Evidence for subaerial emplacement of the Site 1206 lava flow units is derived from their highly vesicular nature, along with the dominance of pahoehoe and a'a flow morphologies (Table T4) (Cas and Wright, 1987; Self et al., 1998). We interpret the hyaloclastite breccia units to be flow-foot breccia (also known as lava deltas) to the overlying lava flows. Flow-foot breccia is formed when lavas flow from land into water (Jones and Nelson, 1970; Moore et al., 1973). The breccia is produced by quenched fragmentation upon contact with water in the nearshore environment, whereas pillowlike lobes are formed where the lava is expelled from tube entries situated below the water level.

We next grouped the lava flow units into 10 "eruption units" (= flow fields) (see Table T3 in the "Explanatory Notes" chapter) by identifying several sets of lava flows that formed by the same eruptions (labeled I-X; Table T4). In general terms, this division follows the consolidation of the volcanic section into eruption units in which the flow-foot breccias are grouped with the overlying lava units (e.g., Units 1-2, 7-8, 10-11, and 13-14) (Tables T3, T4). The basis for this pairing of the basement units is that the hyaloclastite breccia and the overlying lava flows are consistently present as a characteristic lithofacies association situated above the main sedimentary horizons in the section. Also, the paired basement units that form each eruption unit have similar petrographic characteristics. Moreover, analyzed samples from eruption Units I, II, and IV have similar Ti/Zr ratios (see "Geochemistry") (Fig. F3). However, analyzed samples from Units 7 and 8, which tentatively are grouped together as eruption Unit V, differ significantly in chemical composition. Only one onboard sample was taken from each unit for thin section and chemical analysis, and the Unit 7 sample was taken from Section 197-1206A-19R-3 near the top of the unit. Therefore, at this stage it is not certain if basement Units 7 and 8 represent one or more eruption units.

Below, we describe the characteristic architecture of the lava flow units at Site 1206 and briefly discuss the volcanologic implications of the observed lithofacies associations. The terminology used in this section to describe the lavas is outlined in "Physical Volcanology and Igneous Petrology" in the "Explanatory Notes" chapter.

Lava Flow Units

Using the type of flow or lobe contacts along with arrangement of internal flow structures, we identified three lava flow types and one subtype in the Site 1206 basement succession (Table T4). The Site 1206 lavas are categorized as compound pahoehoe (five units), flow-foot breccia (four units), a'a (four units), and transitional (one unit). Because of poor recovery, the flow type of Unit 13 (Section 197-1206A-31R-1) could not be determined with certainty, but the vesicle fabric of the lava appears to be consistent with a pahoehoe flow origin. The key features of the lava flow units are listed in Table T4, and the main characteristics of each lava type are summarized below.

Compound Pahoehoe (Units 1, 4, 6, 7, 10, and, Possibly, 13)

At Site 1206 the compound pahoehoe lavas account for 47% of the total thickness of the volcanic succession and are the thickest subaerial eruption units in the succession, varying from 16 to 39 m (Tables T4, T6). Each unit consists of 5 to >14 vertically stacked lobes ranging from 0.1 to 7.7 m in thickness.

The pahoehoe lobes are characterized by smooth glassy lobe margins grading sharply to aphanitic (hypohyaline to hypocystalline) groundmass, whereas the lobe interiors are typically fine grained and holocrystalline. The lobes have highly vesicular upper and lower crusts and moderately vesicular to nonvesicular lobe interiors (Table T4). Smaller lobes are typically vesicular throughout. The vesicle fabric is isotropic (i.e., shows no evidence of viscous deformation of the lava during flow) and is consistent with laminar flow of low viscosity lava beneath a stationary insulating crust. These lobes also feature segregation structures, such as small pipe vesicles, vesicle cylinders, and megavesicles, which are indicative of lava inflation during emplacement (Self et al., 1998).

Flow-Foot Breccia (Sheet Lobes; Subunits 2a-2c, 8a-8o, 11a-11ae,
and 14a-14d)

These volcanic units consist of very poorly sorted and clast-supported hyaloclastite lapilli breccia intercalated with up to 15 coherent lava intervals. Each lava interval is composed of one to three lava lobes with smooth glassy margins, and the vesicularity of the lobes is variable, typically between 30% and 40% near the top of each breccia unit and decreasing steadily to <3% at the base. This change in vesicularity is consistent with transformation from pahoehoe to pillow morphologies (Self et al., 1998). The lapilli breccia units are always capped by a compound pahoehoe lava and invariably have a conformable depositional contact with the underlying sedimentary units (Table T4; Fig. F4).

The breccia consists of highly angular and sparsely to moderately vesicular (1%-20%) basalt glass fragments, 2-15 mm in size, which are altered to bright green to dark green clay. The basalt glass fragments are well preserved and are characterized by blocky shapes, with straight to highly irregular and convoluted outlines (Fig. F5). Present in the lapilli breccia are larger (20-70 mm) angular clasts of aphanitic lava that have the same lithology as the basalt glass clast population, except for a higher degree of crystallinity. Many of the lava clasts have unaltered glassy lobe margins along one or more of their edges, indicating that they were formed by disintegration of small lava lobes. The breccia has a matrix of black clay (after silt-sized fractions of glassy basalt fragments?). The matrix often contains green clay pseudomorphs of sand-sized basaltic glass shards that exhibit blocky and splinterlike morphologies.

The overall characteristics (or architecture) of the hyaloclastite lapilli breccia units, volcaniclastic texture, and lithofacies associations suggests that they were formed as flow-foot breccia when subaerial lava flowed from land into water.

A'a Lavas (Units 5, 17, and 21)

Three lava units at Site 1206 were identified as a'a flows and account for ~17% of the total lava thickness in the succession (Table T4). The a'a lavas range in thickness from 7 to 17 m (average = 8.4 m). Units 5 and 21 consist of a single lobe (or cooling unit), whereas Unit 17 is composed of two lobes. These lavas are characterized by massive interiors enveloped by thin flow-top and flow-base breccias (0.1-2.3 m thick) consisting of spinose and clinkerlike lapilli-sized basalt clasts (Fig. F6). However, the flow-top breccia of Unit 5 was not recovered and the basal breccia of Unit 21 contains an ash-sized matrix of splinterlike and blocky fragments, indicating fragmentation by contact with water. Internally, these lavas are characterized by a central region with moderately developed viscous flow fabric that is enveloped by zones showing distinct viscous flow fabric (Fig. F6). This fabric is expressed by a subparallel arrangement of flow bands (present as 1- to 2-mm-thick dark gray and subhorizontal wispy bands) and by elongate vesicles drawn out in response to viscous flow at high strain rates. The observed flow fabric is best explained by a sheared laminar flow around a flow center that is transported within a viscous flow as a plug.

Transitional Lava (Unit 18)

Unit 18 is a 3.1-m-thick lava unit that has a pahoehoe base and is capped by a deeply weathered scoriaceous flow top. It features well-developed viscous flow fabric and is the only lava at Site 1206 identified as a transitional type (Table T4). This low frequency of transitional flow type occurrence at Site 1206 contrasts strongly with that of Site 1205 at Nintoku Seamount, where transitional lavas are the most common flow type (see "Physical Volcanology and Igneous Petrology" in the "Site 1205" chapter). The reason for this difference is not obvious but may indicate gentler slopes at Site 1206 compared to those of Site 1205.

Remarks

The pahoehoe lavas and their subaqueous derivatives make up >80% of the lava succession at Site 1206 and represent the low-viscosity end-member of the lava types in the section. These lavas exhibit internal flow architecture similar to that found in modern pahoehoe flow fields in Hawaii and at other Emperor Seamounts (e.g., Hon et al., 1994) (see also the "Physical Volcanology and Igneous Petrology" sections in each site chapter), implying that endogenous growth was an important mechanism for dispersing lava from the source vents at Koko Seamount. We envision that the lava was transported in insulating pathways from the source vents to the active flow fronts, which were nearshore entries similar to those depicted on Figures F7 and F8. As the lava broke out from its pathway, it advanced onward into the sea, where it was subjected to quench fragmentation, which contributed to the growth of the flow-foot breccia (see also Fig. F56 in the "Leg 197 Summary" chapter).

The a'a lavas of Site 1206 represent flows that had the highest bulk viscosity when they were emplaced, which indicates thermally inefficient lava emplacement in open lava channels rather than higher initial melt viscosities (see "Physical Volcanology and Igneous Petrology" in the "Site 1205" chapter).

Petrography

Silicate Mineralogy

The lavas recovered from Site 1206 are all aphyric or olivine-phyric to olivine-plagioclase-clinopyroxene-phyric basalt. Olivine is the most abundant phenocryst phase (up to 10% of rocks examined in thin section). Olivine is observed in both the phenocryst (0.8-5 mm size) and microphenocryst (0.1-0.8 mm size) phases. Euhedral olivine phenocrysts are typically altered to brown clay, serpentine, or green clay, with iddingsite or Fe oxyhydroxide alteration concentrated along the rims or fractures (e.g., Figs. F9, F10, F11, F12). Rarely, they are replaced by calcite or zeolite. In some cases an unaltered interior remains (e.g., Figs. F13, F14, F15). Embayed margins and hollow centers (infrequently observed) are evidence for disequilibrium of the phenocryst phase with the groundmass. Often, Cr spinel inclusions (<0.1 mm in size) are observed in the olivine phenocrysts (e.g., Figs. F16, F17).

Plagioclase phenocrysts constitute up to 4 modal% of some rocks. These are present as subhedral laths that average ~0.7 mm in length. They often exhibit optical zonation, fractures, resorbed margins, and embayments, indicating resorption (e.g., Fig. F18). Frequently, plagioclase forms glomerocrysts with clinopyroxene and, rarely, olivine (e.g., Figs. F19, F20, F21).

Where present, clinopyroxene phenocrysts account for up to 3 modal% of some rocks. These are subhedral and 0.5-2 mm in size. They commonly are incorporated in glomerocrysts with plagioclase. Rare Cr spinel phenocrysts (average size = 0.4 mm) are present in samples from Units 6 and 17 (e.g., Fig. F22).

Many basalt samples exhibit a porphyritic texture, composed of larger phenocrysts in a fine-grained to aphanitic groundmass of plagioclase, clinopyroxene, glass, titanomagnetite, and, occasionally, olivine. Sometimes the groundmass is characterized by a subophitic texture, but most commonly it has an intergranular to intersertal texture.

In the groundmass, plagioclase laths are euhedral to subhedral, averaging ~0.3 mm in length, and are sericitized and/or altered to clay in some samples. In Subunit 2c, groundmass plagioclase laths exhibit swallow-tail terminations, indicative of high cooling rates and rapid growth. These can be found sericitized and altered to clay in some samples. Groundmass clinopyroxene is usually anhedral. Titanomagnetite is a minor constituent of the groundmass, exhibiting small, often octahedral, and sometimes dendritic morphologies (<0.5 mm in size) that are sometimes partially altered to maghemite (e.g., Fig. F23; see also below). Groundmass olivine is rare, but when present it is completely altered to iddingsite or green clay. In some instances (such as Section 197-1206A-19R-3), groundmass olivine may actually be a microphenocryst phase (e.g., Fig. F24). In other areas, olivine displays a seriate texture. Glass is usually devitrified and completely altered to brown clay, green clay, or zeolite. Vesicles are variable in abundance (up to 20%). Some are surrounded by segregated material, often composed of plumose altered clinopyroxene and titanomagnetite crystals that radiate out from skeletal plagioclase (e.g., Fig. F25).

Basalt recovered from Site 1206 differs from that recovered from Site 1205 in the lower abundance of groundmass olivine. The overall texture of the groundmass also differs, from subtrachytic in basalt from Site 1205 to generally intersertal to subophitic in that from Site 1206. In contrast to basalt from Sites 1203 and 1205, where plagioclase is the dominant phenocryst phase, basalt from Site 1206 is dominantly olivine-phyric.

Opaque Mineralogy

The opaque mineralogy of the Site 1206 basalt units is dominated by titanomagnetite. Primary sulfide is notably absent from every thin section examined from this site. Secondary pyrite was observed in only two instances (Units 7 and 21) (see Table T7). In general, titanomagnetite has a morphology that is predominantly a mixture of subhedral skeletal octahedra and dendritic forms (Fig. F26), but with a few units exhibiting only subhedral skeletal octahedral forms (e.g., Units 5 and 21) (Fig. F27). Titanomagnetite exhibits varying degrees of oxidation (see Table T7); ilmenite oxidation lamellae are present in the titanomagnetite throughout the basalt units (e.g., Fig. F28). The development of ilmenite oxidation lamellae is occasionally accompanied by alteration to maghemite along cleavage planes, fractures, and at the crystal margins (e.g., Figs. F28A, F29). In basalt units containing maghemite, it is common that many titanomagnetite crystals have escaped this alteration (e.g., Fig. F30). As we note in other site reports from Leg 197, maghemite and titanomaghemite (or Ti-bearing maghemite) are indistinguishable using reflected-light microscopy.

Cr spinel is also found in all lava flow units except Units 7, 8, 10, 18b, and 21. The Cr spinel is usually found as inclusions in olivine phenocrysts (e.g., Figs. F31, F32) and may exhibit titanomagnetite overgrowths (Fig. F33). Cr spinel is present in the olivine-rich regions of Units 6 and 17 as both inclusions in olivine and as discrete phenocrysts. In Unit 6, the Cr spinel phenocrysts are euhedral (Fig. F33) with a small titanomagnetite overgrowth (Fig. F34). The titanomagnetite overgrowths on Cr spinel contain ilmenite oxidation lamellae and maghemite development along fractures, but these features are truncated by the Cr spinel (Fig. F35). Olivine in Unit 6 is highlighted by a rim of Fe oxyhydroxide-rich iddingsite (Fig. F36A), a reddish brown substance consisting of smectite, chlorite, and goethite + hematite (e.g., Deer et al., 1992). In Unit 17, olivine is highlighted by a rim of serpentine and the Cr spinel phenocrysts are round (Fig. F36B). The morphology of the Cr spinel phenocrysts is due to extensive overgrowths of and complete replacement by titanomagnetite (Fig. F37). The titanomagnetite has subsequently been oxidized and exhibits ilmenite oxidation lamellae (Fig. F37C, F37D). A similar relationship between titanomagnetite and Cr spinel is seen in Unit 5 (Sample 197-1206A-15R-1 [Piece 7A, 72-74 cm]) and suggests discrete Cr spinel phenocrysts may also have been present in this unit but have subsequently been replaced by titanomagnetite (cf. Figs. F37A, F37B, F38). Cr spinel inclusions in olivine from Unit 6 are usually replaced by titanomagnetite, especially if they lie in the iddingsite alteration rim (Fig. F39).

Plagioclase and clinopyroxene in the glomerocrysts present in Unit 5 also contain Cr-rich opaque minerals (Fig. F40). These may be small chromite crystals, as they are dull gray in reflected light compared to titanomagnetite and lack the characteristic blue-gray color of Cr spinel.

Geochemistry

Major and trace element abundances were determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) (see "Physical Volcanology and Igneous Petrology" in the "Explanatory Notes" chapter) for 20 basaltic samples from the basement rocks at Site 1206 (Table T8). In the total alkalis vs. SiO2 classification plot only three of the 20 samples are alkalic basalt (basalt from Units 8 and 18 and one of the three samples from Unit 4), whereas all other samples plot in the tholeiitic basalt field (Fig. F41). The alkalic nature of basalt from Units 8 and 18 is confirmed by their relatively high TiO2 and Zr content at a given MgO content; however, the compositionally heterogeneous samples of Unit 4 do not have equally high TiO2 and Zr contents (Fig. F42), so the lone sample from Unit 4 that plots in the alkalic field may have experienced K2O addition during alteration. At 7 wt% MgO, Site 1206 samples range significantly in TiO2 and Zr content; clearly, the lavas do not define a single liquid line of descent. Apparently, the lavas recovered at Site 1206 evolved from a continuum of parental magmas ranging from tholeiitic basalt with relatively high abundances of incompatible elements to slightly alkalic basalt with higher abundances of incompatible elements.

Basalt from Koko, Nintoku, and Suiko Seamounts define inverse trends in plots of Al2O3, TiO2, Na2O, K2O, and Zr abundance vs. MgO content (Fig. F42). Samples from Koko and Suiko Seamounts overlap in MgO variation plots. In contrast, many of the alkalic lavas at Nintoku Seamount have lower MgO contents and they show trends to lower CaO and Sc contents (Figs. F42, F43), indicative of clinopyroxene fractionation. Such trends are typical of the alkalic lavas erupted during the postshield growth stage of Hawaiian volcanoes (see "Physical Volcanology and Igneous Petrology" in the "Site 1205" chapter). Two of the three Site 1206 samples (one from Unit 4 and one from Unit 8) that plot in the alkalic field (Fig. F41) also have relatively low CaO contents, but they do not have low abundances of Sc (Fig. F42; Table T8).

Abundances of incompatible trace elements such as Ba, Sr, and Zr further demonstrate the geochemical similarity of basalt from Koko and Suiko Seamounts and the higher abundance of Ba and Sr in basalt from Nintoku Seamount (Fig. F44). The alkalic basalt from Koko Seamount does not have the high Sr and especially the high Ba content that characterizes the alkalic basalt from Nintoku Seamount (Fig. F44). In contrast to the differences in Sr and Ba, basalt from all three seamounts define the same Ti vs. Zr trend (Fig. F44).

An important aspect of the growth of most Hawaiian volcanoes is the gradual temporal trend from dominantly tholeiitic basalt during the shield stage followed by dominantly alkalic basalt during the postshield stage (Clague and Dalrymple, 1987). The dominance of alkalic lavas at Site 1205 on Nintoku Seamount (Fig. F45) clearly shows that the basement represents the postshield stage, whereas at Suiko Seamount only the uppermost three lava units are clearly alkalic basalt (Jackson et al., 1980). Like lavas from Hole 433C at Suiko Seamount, basalt from Site 1206 at Koko Seamount is dominantly tholeiitic (Fig. F45). We infer that at Site 1206 the late shield-stage of growth at Koko Seamount was sampled, but lavas that likely represent the postshield and rejuvenated stages have been recovered by dredging (Clague and Dalrymple, 1973, 1987).

As discussed in "Physical Volcanology", the 22 lithologic units (Fig. F2) can be grouped into 10 eruption units and 7 intercalated sediment units (Table T3). Multiple samples were analyzed from eruption Units I, II, and IV, but recovery was low in the intervals represented by these units, especially from Cores 197-1206A-4R to 6R, 9R to 15R, and 16R to 17R (Fig. F2). Our goal was to geochemically evaluate the interpretation that these intervals might contain more than one eruption unit. Abundances of Ti and Zr are relatively immobile during postmagmatic alteration; hence, the Ti/Zr abundance ratio can be used to evaluate the grouping of lithologic units into eruption units. The Ti/Zr ratio clearly distinguishes eruption Units I, II, and IV (Fig. F3). However, samples analyzed from eruption Units V and VI have disparate Ti/Zr ratios (Fig. F3). In particular, eruption Unit V contains one of the two distinctly alkaline samples. In this case, the geochemical data are inconsistent with the grouping of lithologic Units 7 and 8 into a single eruption Unit V.

Lavas in eruption Units I, II, and IV contain variable amounts of olivine (Fig. F2), and undoubtedly the variable MgO content of samples from eruption Units I and IV reflects varying amounts of olivine in the analyzed samples (Fig. F46). The compositional variation of eruption Units I and IV in Figure F46 could be explained by addition and subtraction of Fo70 olivine. It is likely, however, that olivine in these lavas has a higher forsterite content; approximately Fo82 is inferred using the whole-rock (wr) Fe/Mg ratio (assuming 90% of Fe is Fe2+ and (Fe/Mg)ol/(Fe/Mg)wr = 0.3). A more realistic explanation of the trend in Figure F46 may be combined fractionation of olivine and plagioclase. Plagioclase is a minor phenocryst phase in these eruption units.

Although the three samples from eruption Unit II have similar Ti/Zr ratios and abundances of Ti, Zr, and MgO (Figs. F3, F46), each sample has distinctive geochemical features. Among these three samples, Sample 197-1206A-8R-1, 99-101 cm, has the highest Al2O3 content (16.1 vs. ~15 wt%), whereas Sample 197-1206A-8R-1, 2-3 cm, has the lowest K2O content (0.55 vs. 0.8-0.9 wt%) and Sample 197-1206A-8R-2, 3-5 cm, has much lower abundances of CaO, Na2O, Sr, and Ba (Table T8). These differences may indicate that this eruption unit was formed by more than one geochemically distinct magma or possibly that these compositional differences reflect the effects of postmagmatic alteration.

NEXT