In general, the fresh dacites from PACMANUS are vesicular and glassy with variable amounts of feldspar microlites. The phenocryst assemblage includes plagioclase, pyroxene, and Ti-magnetite. Based on point counting of 23 thin sections (500–800 points each; data documented in Binns, Barriga, Miller, et al., 2002), three types of dacite can be distinguished: moderately porphyritic dacite (2%–3% plagioclase, 1% pyroxene, 0.5% Ti-magnetite), sparsely porphyritic dacite (0.5%–2% plagioclase, 0.1%–1% pyroxene; <0.1%–0.5% Ti-magnetite), and aphyric dacite (0.5% phenocrysts). Although determined attempt was made to evaluate each piece of fresh volcanic rubble for indications of stratigraphic relevance (number of flows or cooling units), no such correlation was deemed possible. From Sites 1188, 1189, and 1190, only a few pieces of relatively fresh material were recovered (<50 smaller than fist–sized pieces in total) so only subtle inferences regarding recent volcanic stratigraphy are possible. More fresh dacite pieces (68) were recovered from Site 1191. However, even in pieces long enough to be considered oriented (25% of pieces recovered, of which only half were convincingly long enough to have escaped rotation in the core barrel) closely adjacent pieces show radically different orientations in vesicle flattening and alignment, suggesting most if not all recovered material was from disaggregated lava lobes.

Site 1188 (Snowcap)

Megascopic Features

The uppermost section sampled at Site 1188 (Cores 193-1188A-2R through -4R; Core 1R had no recovery) consists of 15 curated pieces of volcanic rock fragments. Fewer than half of the fragments are more than a few centimeters in longest dimension. The core is moderately vesicular plagioclase- and clinopyroxene-phyric dacite. In pieces long enough to suggest they are oriented, vesicles commonly show elongation (cigar-shaped in three dimensions rather than merely flattened) with a typically normal orientation with respect to the long axis of core, indicating horizontal flow on the seafloor; therefore, an intact lava flow was likely sampled. In the top of Core 193-1188A-7R, beneath an interval of aphyric, pervasively altered dacite, three pieces of fresh dacite were recovered. In all megascopic and microscopic features, these pieces appear identical to the fragments in Cores 193-1188A-2R through 4R, and we can find no reason to contend the interpretation of shipboard scientists (Shipboard Scientific Party, 2002b) that these were pieces of rubble from the upper part of the section that fell into the hole and were resampled during the coring operation. Therefore, even with the low recovery, we concur with the shipboard interpretation that the unaltered, morphologically youthful part of the lava section at this site is no more than ~35 m thick.


Plagioclase and pyroxene (predominantly clinopyroxene with minor orthopyroxene) are euhedral to subhedral (Fig. F3A) and <2 mm long, with rare exceptions. Twinned feldspar grains are common (Fig. F3B), but oscillatory zoning is less common. Only rare evident quench textures or skeletal growth features suggest little radical undercooling of the lava. Melt inclusions are common in plagioclase (Fig. F3C) and more rare in pyroxene. Pyroxene forms elongate prisms (Fig. F3A), commonly broken but with euhedral faces. Small subhedral magnetite phenocrysts are rare (Fig. F3D) and commonly occur proximal to plagioclase phenocrysts. Magnetite microphenocrysts <0.2 mm in length are common as inclusions in plagioclase and pyroxene.


Groundmass in virtually all thin sections examined is 55%–65% fresh, brown, volcanic glass. The groundmass glass is permeated with 35%–45% acicular plagioclase microlites with generally a 10:1 aspect ratio. Very fine grained pyroxene and opaque minerals are also present. In most thin sections, the groundmass microlites are strongly flow aligned (Fig. F3).


Vesicularity appears to decrease downhole from 10% in Core 193-1188A-2R to 5% in Core 4R. The majority of the vesicles are flattened and flow aligned parallel to microlite fabric in the groundmass.

Site 1189 (Roman Ruins)

Megascopic Features

Only the uppermost four pieces of core fragment from Site 1189 are fresh volcanic rocks. These are vesicular but virtually aphyric. The rest of Core 193-1189A-1R and the upper four pieces of Core 2R are moderately altered but are aphyric to very sparsely plagioclase phyric dacite.


Very rare, rounded microphenocrysts of feldspar occur in the moderately altered samples from Cores 193-1189A-1R and 2R and in similarly altered samples from deeper in the hole (Fig. F4A). Considering the rounded plagioclase morphology and the observation that some feldspars exhibit normal zoning and some show reverse zoning suggests that plagioclase was on the liquidus during emplacement of the dacite at Roman Ruins.

Groundmass and Vesicles

Fresh, brown glass constitutes the groundmass of the upper few pieces of core from Site 1189. In the freshest samples, 15%–25% plagioclase microlites show no preferred orientation despite the flattened and aligned vesicles (30% by volume of the rock). In one of the least altered samples from Core 193-1189A-2R, vesicles are rounded and plagioclase microlites have a distinct jackstraw distribution (Fig. F4B). The much more extensive vesiculation in the upper part of Core 193-1189A-1R, the distinct change in alteration character and vesicle morphology, and the change from complete absence to rare occurrence of detectable phenocrysts suggest that the lavas below Core 193-1189A-1R could be a separate flow unit.

Site 1190 (Reference Site)

Megascopic Features

Samples of fresh volcanic rock from the reference site are moderately vesicular and sparsely plagioclase and pyroxene phyric (Fig. F5A). Three holes were drilled, with no recovery from the first, only one cored interval from the second, and a total of just over 13 m penetration at the third with <25 fragments of volcanic rock recovered. None of these pieces were large enough to be considered oriented with respect to the way up. The samples are all sparsely plagioclase phyric, and we can detect no features to depart from the shipboard interpretation that all the pieces of rock recovered are from a single flow. Because the pieces have no distinguishing characteristics and none of the pieces are oriented, we have no criteria from which to infer whether or not the pieces are from a single flow or are part of pile of volcanic rubble.


Plagioclase (virtually ubiquitously 2% by volume) is generally euhedral to subhedral and rarely exhibits zoning, but inclusions of magnetite and melt are common. Subtle rounding of phenocryst margins is common. Pyroxene is much less abundant than plagioclase and occurs as subhedral prisms with rounded corners and crystal edges. Magnetite phenocrysts are smaller than plagioclase or pyroxene but distinctly larger than groundmass phases. Glomerocrysts of plagioclase, pyroxene, and magnetite are common, and the groundmass glass surrounding these clusters of crystals is free of microlites (Fig. F5A). A discussion of this phenomenon can be found in the Site 1190 chapter of the Leg 193 Initial Reports volume (Shipboard Scientific Party, 2002d).

Groundmass and Vesicles

Groundmass glass contains abundant acicular plagioclase microlites with fewer opaque grains and distinct flow banding manifested as dark and lighter wisps (Fig. F5B). The darker bands are more voluminous and have gradational to sharp boundaries with the lighter bands. The dark bands contain abundant microvesicles. The lighter bands have more abundant volcanic glass. This agrees with shipboard observations (Shipboard Scientific Party, 2002d), although shipboard samples were described with more abundant plagioclase microlites in the darker bands as well. The bands were interpreted to suggest that phase separation may have been favored where microlites were more abundant. However, even in flow banded areas, radial clusters of microlites are common, with some microlite alignment present. We also observe that the most striking difference between the bands is the proportion of glass to microvesicles and not a marked difference in plagioclase microlite abundance. Vesicle abundance ranges from 5% to 20%, and whereas most samples are characterized by flattened and elongated vesicles, a few (i.e., Sample 193-1190C-2R-1, 10–12 cm) have a distinctly bimodal vesicle distribution with small (0.2 mm) spherical and larger, >1 mm, flattened and aligned vesicles.

Site 1191 (Satanic Mills)

Megascopic Features

Fresh vesicular aphyric to sparsely plagioclase phyric dacite was recovered from the subsurface of the Satanic Mills hydrothermal site. Drilling penetrated ~20 mbsf. The most striking feature of the samples recovered from this location is the transition from fresh to incipiently altered to moderately altered rhyodacite within a few centimeters of the seafloor. The intensity of alteration a this site is lower relative to the pervasive alteration reported from deeper intervals at Sites 1188 and 1189 (Shipboard Scientific Party, 2002b, 2002c). Although a few pieces of fresh volcanic rock appear in deeper cores, because there is no morphologic or petrologic criteria to distinguish these as more recent flows, we interpret these as pieces of fresh material falling into the borehole during coring.


Many thin sections are aphyric; however, subhedral to euhedral plagioclase is the most common phenocryst phase but rarely exceeds 1 mm in long dimension (Fig. F6). Rare pyroxene phenocrysts are less elongate, although they still appear slightly rounded. Magnetite phenocrysts are rare.

Groundmass and Vesicles

The groundmass from all fresh samples examined from Site 1191 comprise brown volcanic glass and fine, disseminated plagioclase microlites. The microlites commonly exhibit flow alignment. Vesicles are fairly abundant, exceeding 30% by volume in some samples. Several pieces have a distinctly bimodal vesicle distribution, with one set of small (average = 0.2 mm) spherical vesicles and another set of larger (>1 mm) flattened and aligned vesicles. With depth, there seems to be a decrease in the abundance of flattened and aligned vesicles, but the bimodal size distribution remains.

Bulk Rock Chemistry

Data from nearly all samples plot near the dacite/rhyolite boundary in a total alkali vs. silica diagram (Fig. F7), and compositions (Table T1) are consistent with the field of bulk rock analyses for dredge samples from Pual Ridge compiled by Yang and Scott (2002). The data indicate lavas from Roman Ruins (Site 1189) are the most mafic (e.g., 64 wt% SiO2), whereas samples from the other sites were all more silica rich.

The uppermost lavas from Site 1189 have MgO of 1.8 wt% and elevated CaO relative to fresh lavas from other sites (Fig. F8). One of the rare, relatively fresh dacites sampled from ~130 mbsf at Site 1189 has an elevated MgO (1.3 wt%) as compared to the other sites sampled (0.7–0.9 wt%) but most closely resembles lavas from Sites 1188 and 1190 in terms of most major element oxides. Covariation diagrams of SiO2 relative to other major element oxides (MgO, Al2O3, Fe2O3, CaO, and alkalis) exhibit consistent fractionation trends from the most primitive samples (Site 1189) to the most evolved (Site 1190).

Plots of TiO2 vs. Zr and SiO2 vs. Ti/Zr (Fig. F9) also illustrate the geochemical variability present in the fresh lavas from Pual Ridge, indicative of magmatic differentiation. Ba increases from 250 to >400 ppm, consistent with magmatic differentiation, although relatively high Ba (>500 ppm) may be indicative of incipient alteration. Chondrite-normalized REE compositions from all sites (N = 28) are indistinguishable, indicating derivation from a common parent (Fig. F10). The REE patterns of fresh dacites from the subsurface at PACMANUS show negative Eu anomalies with Eu/Eu* values ranging from 0.98 to 0.69. A slight negative correlation between Eu/Eu* and SiO2 defined by most samples (Fig. F11) indicates the development of the Eu anomaly was probably associated with igneous fractionation processes. A weak positive correlation between Eu/Eu* and CaO and a more well-developed positive correlation between Eu/Eu* and Sr (Fig. F11) indicates plagioclase fractionation was an important process during the petrogenetic evolution of the dacitic magmas from Pual Ridge. A summary of key petrologic and geochemical data is in Table T3.

Mineral Chemistry

Plagioclase compositions from fresh Site 1188 (Snowcap) lavas range from An37 to An44. Clinopyroxene is Wo40-En40-Fs20, and orthopyroxene is Wo3.5-En55-Fs41.5 (Table T3; Fig. F12). At Roman Ruins (Site 1189) the rare plagioclase phenocrysts are the most calcic plagioclase compositions analyzed from the Leg 193 fresh lava sample set, ranging from An49 to An52. Plagioclase analyses from altered samples recovered from deeper at Site 1189 range from An40 to An50. Plagioclase compositions from Site 1190 (reference site) lavas are An43 to An47, somewhat less calcic than the nearby Site 1189 lava but significantly more calcic than plagioclase sampled from Snowcap or Satanic Mills. Fresh lava from Site 1190 has orthopyroxene with the same composition as the Snowcap site, but clinopyroxene is slightly more calcic (Wo43-En38-Fs19). The Satanic Mills (Site 1191) plagioclase phenocrysts contain the lowest anorthite content (An35 to An40) measured in fresh lavas sampled during Leg 193. Clinopyroxene compositions appear identical to those from nearby Snowcap, but orthopyroxene (Wo3.7-En62-Fs34.3) contains lower iron and higher magnesium than samples from other locations.