Atlas of alteration textures in volcanic glass from the ocean basins

The interaction of sub-seafloor volcanic glass with circulating fluids produces secondary minerals as well as alteration textures that penetrate into the glass (e.g., Thorseth et al., 1995; Fisk et al., 1998a; Alt and Mata, 2000; Furnes et al., 2001a; Josef, 2006). These alteration textures are found in basalts from the flanks of ocean rifts, seamounts, back-arc basins, and marginal seas. The alteration textures include micron-sized etch pits and tunnels that are located at the interface of fresh glass and its alteration products. This petrographic atlas aims to bring together and illustrate the full spectrum of alteration textures in marine lavas to show the variety of alteration textures found in volcanic glass and hyaloclastites collected from the ocean crust. In particular we focus on the size, morphology, distribution, and infilling of granular cavities and tubular tunnels.

A selection of annotated petrographic images from a collection of 119 samples spanning the world’s ocean basins is provided to systematically illustrate the key textural characteristics of glass alteration. A guide and glossary to the principal features is provided and an accompanying classification scheme is given to identify the key morphotypes of glass alteration. This expands on earlier classification schemes (Furnes and Staudigel, 1999; Josef, 2006; Staudigel et al., 2006, 2008; McLoughlin et al., 2009) and identifies several previously unrecognized morphotypes.

The atlas is intended as an illustrated guide for geologists, microbiologists, and astrobiologists studying glass alteration. We realize that as researchers further explore their collections and as more deep-sea environments are sampled, new forms of glass alteration will be found and documented; thus, this guide represents the current state of knowledge. Some alteration textures have previously been argued to represent biological alteration products and trace fossils (e.g., Fisk et al., 1998a; Torsvik et al., 1998; Furnes et al., 2001a, 2001b, 2002, 2008; Furnes and Muehlenbachs, 2003; Banerjee and Muehlenbachs, 2003; Thorseth et al., 2003; McLoughlin et al., 2009; Staudigel et al., 2008); however, this study is not designed to support or refute claims of biogenicity of the alteration of basaltic glass. Also, this work does not investigate the secondary mineralogy of altered glass, referred to as palagonite, a mixture of iron oxyhydroxides and phyllosilicates (Stronick and Schmincke, 2002), and we have not characterized the secondary minerals, such as carbonates, zeolites, and phyllosilicates, that occur in voids and fractures in the glass. These minerals are indicative of conditions in the seafloor and may be useful for correlating conditions of alteration with glass alteration features; however, this is the subject of an ongoing study. If the alteration textures are biotic and if specific textures can be correlated with subsurface conditions, then they could help researchers understand the evolution of the marine subsurface environment from the Archean to the present.

Granular and tubular alteration textures of oceanic volcanic glass have been illustrated in transmitted-light photomicrographs since the 1960s (Morgenstein, 1969). In that first study, glass/palagonite alteration boundaries and linear features in black-and-white photographs were described as “micro-channels” and “hair channels.” More recently, transmitted-light photomicrographs of alteration features in seafloor and sub-seafloor basalt glass have been published by a number of authors (e.g., Giovannoni et al., 1996; Fisk et al., 1998a, 2006; Furnes and Staudigel, 1999; Fisk and Giovannoni, 1999; Christie et al., 2001; Furnes et al., 2001a, 2002; Banerjee and Muehlenbachs, 2003; Storrie-Lombardi and Fisk, 2004; Ivarsson et al., 2008; Staudigel et al., 2008; Cockell and Herrera, 2008; McLoughlin et al., 2009, 2010; Heberling et al., 2010). These studies have, in general, included a limited number of images to illustrate the granular or tubular structures, and they have not documented the full range of alteration textures now known from oceanic igneous glass. An extensive unpublished collection of photomicrographs also exists (Josef, 2006).

Over this more recent period (1996 to the present), alteration features in oceanic basalt glass have also been illustrated in backscattered electron images, transmission electron images, and energy-dispersive X-ray spectroscopy (EDS) maps (e.g., Furnes et al., 1996, 1999; Torsvik et al., 1998; Alt and Mata, 2000; Thorseth et al., 2003; Kruber et al., 2008; Cockell et al., 2009). Also, similar features have been documented with transmitted-light photographs of metamorphosed pillow-lava rims from Archean to Phanerozoic ophiolites (Furnes et al., 2001b, 2004, 2008; Furnes and Muehlenbachs, 2003; Staudigel et al., 2006, 2008) and an Archean mafic tuff (Lepot et al., 2011). Photographs of granular and tubular alteration in basalts from the marine/land transition have also been published (Fisk et al., 2003; Walton and Schiffman, 2003; Walton, 2008; Cousins et al., 2009; Montague et al., 2010). Interestingly, similar transmitted-light photomicrographs of alteration features in pillow lavas erupted into fresh water are not evident in the literature.

Common alteration textures, such as tunnels in volcanic glass, were until recently informally classified by several authors, so synonyms for these textures exist in the literature. A more formal ichnotaxonomic classification was suggested by McLoughlin et al. (2009), which considered potential bioalteration textures as trace fossils and recognized two ichnogenera and five ichnotaxa based on a selection of samples from the in situ oceanic crust and Phanerozoic ophiolites. This atlas expands on these five ichnotaxa or morphotypes, and outlines seven morphological criteria and provides names for features that can help unify discussion of the morphotypes.

The percentage of glass alteration in sub-seafloor basalts can be estimated visually, and the percent of that total alteration that is attributed to biotic versus abiotic processes has been derived by point counting of thin sections from the Mid-Atlantic Ridge, the Costa Rica Rift, and Lau Basin (Furnes et al., 2001a). From 2% to 60% of the glass was altered, with about half of this alteration being granular and tubular and the remainder being “abiotic” secondary minerals. The amount of granular and tubular alteration has also been visually estimated in basalt glass at the marine/land transition (Cousins et al., 2009; Montague et al., 2010). One of these studies (Cousins et al., 2009) from the glacial/marine transition of James Ross Island, Antarctica, found that the samples exposed to seawater tended to have more granular and tubular alteration than samples exposed to fresh water. Alteration of a Hawaiian subsurface hyaloclastite was indexed from 1 to 6, with 1 being no glass alteration to 6 being complete alteration (Montague et al., 2010). Indices were mostly 2–3. From these three studies, it appears that granular and tubular alteration is ubiquitous and more abundant in marine water than fresh water.

Although most photographic documentation of tunnels has come from basalts, there are examples of tunnels from other silicates. There are two examples from felsic rocks—one of these is from a rhyolite tuff from central Oregon (United States) (Fisk et al., 1998b) and the other is from a submarine clastic tuff from the western Pacific (Banerjee and Muehlenbachs, 2003). Also, tunnels have been documented in olivine from an olivine basalt collected from the marine/land transition in Hawaii and in dunites from central Oregon and northern California (Fisk et al., 2006).

It has been hypothesized based on several lines of evidence that some of the tunnel and granular alteration features are produced biotically. In support of this, biological staining has revealed that nucleic acids can be found at the interface of fresh and altered glass near tubular and granular textures and in some tubular forms (e.g., Furnes et al., 1996; Giovannoni et al., 1996; Torsvik et al., 1998; Banerjee and Muehlenbachs, 2003). It has been shown theoretically that basaltic glass can yield sufficient energy to support chemolithoautotrophic growth (Bach and Edwards, 2003), and culture-independent sequencing studies have shown that the microbial population inhabiting the sub-seafloor is distinct from that found in both overlying seawater and seafloor sediments and is up to 4 times larger (Mason et al., 2009; Santelli et al., 2008). Controlled laboratory experiments have found that enhanced, localized dissolution occurs in volcanic glass inoculated with microorganisms, relative to abiotic controls (Thorseth et al., 1995; Staudigel et al., 1995). Comparative analysis of pillow-basalt rims and interiors suggests that biological activity has lowered the δ¹³C of the rim relative to the basalt interior (e.g., Furnes et al., 2001c). Partially fossilized, mineral-encrusted microbial cells have been observed in or near etch pits on altered glass surfaces, and these pits have forms and sizes resembling the associated microbes suggesting that the microbes are involved in pit formation (Thorseth et al., 1992, 2001, 2003). Although these studies support the hypothesis of biologically mediated tunnels, it has not yet been possible to cultivate microorganisms that create tunnel shapes, and abiotic mechanisms of tunnel production have been proposed such as for the Archean mafic tuffs, which experienced conditions very different from basalts in our collection (e.g., Lepot et al., 2011). So although the weight of evidence is in favor of the biological formation of complex tunnels, the question has not been answered.

In attempts to understand the origin of the alteration textures, several geochemical tools have been used to examine the contents of tunnels and the chemistry of the surrounding glass and alteration products. These studies have included: electron probe microanalysis (Furnes et al., 1996; Torsvik, et al., 1998; Storrie-Lombardi and Fisk, 2004); scanning and transmission electron microscopy (Alt and Mata, 2000; Thorseth et al., 2003; Benzerara et al., 2007; McLoughlin et al., 2011; Knowles et al., 2012); Raman and/or infrared spectroscopy (Preston et al., 2011); and synchrotron-based X-ray microprobe techniques (Benzerara et al., 2007; Staudigel et al., 2008; Knowles et al., 2011, 2012; Fliegel et al., 2012). It has been hypothesized that Fe(II) is an energy source for microbial metabolism, and electron microprobe analyses of palagonite near “biotic” alteration has higher Fe than palagonite near “abiotic” alteration (Storrie-Lombardi and Fisk, 2004). Analysis of the 100–300-nm-wide rim of a tunnel in glass shows that the glass lost Fe (Alt and Mata, 2000) and this loss of Fe is consistent with a gain in Fe in the alteration material near “biotic” alteration (Storrie-Lombardi and Fisk, 2004). Transmission electron microscopy and synchrotron-based X-ray microprobe analysis show the presence of partially oxidized Fe (Benzerara et al., 2007; Knowles et al., 2011; Fliegel et al., 2012) and organic carbon in tunnel-filling smectite (Benzerara et al., 2007). The X-ray microprobe analysis also shows that tunnels are produced by the dissolution of the glass. Nano-Secondary Ion Mass Spectrometry analyses of carbon, nitrogen, and manganese associated with micropores in glass suggest that these are remnants of manganese-oxidizing bacteria (McLoughlin et al., 2011). Raman spectroscopy indicates that tunnels contain complex organic compounds such as amides and esters, which could be left by microbial inhabitants (Preston et al., 2011). These studies show that microbes and microbial processes are associated with some of the alteration features described in this paper.

The biogeochemical controls on the abundance, distribution, and diversity of alteration textures in volcanic glass, however, are yet to be identified, and we hope that the framework presented herein will aid future investigations of the these controls on (bio)alteration. In addition, this collection of alteration textures may be an informative companion for those studying the alteration of (meta)volcanic glass in ophiolites and/or Precambrian greenstone belts and provide a context for interpreting proposed trace fossils that are hypothesized to represent some of the earliest evidence for life on Earth (Furnes et al., 2004). Likewise, for astrobiologists who may one day study alteration in igneous rocks from other planetary bodies (cf. Fisk et al., 2006), this will provide a terrestrial reference frame for the range of currently known textures.

Samples for this study are from existing collections and are primarily cored sub-seafloor igneous rocks that were collected and archived by the Deep Sea Drilling Project (DSDP) and the Ocean Drilling Program (ODP). The location of volcanic glass within the archived cores was determined by first reviewing Initial Reports volumes of the Deep Sea Drilling Project and the Initial Results volumes of the Ocean Drilling Program. Then, during visits to the sample repositories at Scripps Institution of Oceanography, Texas A&M University, and Lamont-Doherty Earth Observatory, the presence of glass was verified visually and samples were collected from the working halves of the cores. This collection of DSDP and ODP samples was supplemented with a small number of samples from an Integrated Ocean Drilling Program (IODP) expedition as well as from seafloor outcrops that were collected by submersibles. In total, the samples come from 21 DSDP expeditions, 15 ODP and IODP expeditions, and 5 manned and unmanned submersible expeditions. Samples are from the Pacific, Indian, and Atlantic Oceans, the Mediterranean Sea, and some adjacent seas. Basalts from ocean rifts, seamounts, and back-arc spreading ridges are included in the study. The cored samples come from a range of depths into the volcanic basement, ranging from the sediment/basalt contact (<0.5 m into basement, mib) to 320 mib. Samples collected by submersible are from outcrops on the seafloor. The samples are primarily the rims of pillow basalts, sheet-flow margins, and interflow breccias. The youngest cored basalt examined was <0.4 Ma and the oldest was 167 Ma. At some cored sites, samples from multiple depths were examined. Standard 26 mm × 46 mm polished petrographic thin sections were examined. These are listed in Table 1, and Figure 1 shows the global distribution of samples. All samples, except 482D 11R2 32, which appears to have been at ≤150 °C at the time of collection (Duennebier and Blackinton, 1980), were from the seafloor or shallow subsurface where ambient temperatures were compatible with life (<100 °C).

The petrographic thin sections listed in Table 1 were examined with a petrographic microscope fitted with objective lenses with 4×, 10×, 20×, and 40× magnification. Thin sections are nominally 30 μm thick, which is 3–30 times the diameter of most tunnels and thus permits viewing tunnels in three dimensions by changing the elevation of the microscope stage. The observation strategy was to survey the whole thin section with the 4× objective lens to locate areas of glass. These glasses were then viewed at 10× and/or 20× to locate regions that contained altered glass and glass-alteration textures. Glass commonly was limited to less than 25% of the thin section area and most regions of altered glass were examined at one of these two higher magnifications. The alteration was photographed at 40× and sometimes at 10× for larger features or to show the context of the feature being illustrated. All of the morphotypes of alteration found in our selection of thin sections are illustrated in the figures.

The optical images shown here were obtained using a Nikon LV100Pol polarizing microscope at the Centre for Geobiology in Bergen, Norway, and photographed using an DS-Fi1 color camera with 5.24-megapixel resolution coupled to NIS-Elements BR 2.30 software. The images were saved in Joint Photographic Experts Group (.jpg) format (2560 × 1920 pixels).

The alteration features are summarized in line drawings in Figure 2, and each line drawing references a photograph that illustrates the feature (Figs. 3–31). The diversity of alteration features can be illustrated with a subset of 26 thin sections (identified by bold italics font in Table 1). Each photograph is labeled with the sample identification, a scale bar, and descriptive terms. Also the glass and major features such as fractures, vesicles, and minerals are annotated.

Features are first separated into major categories of granular and tubular forms (Fig. 2, upper left panel) as previously described (Furnes et al., 2001a). In addition to the granular form, a bud-shaped form is recognized here that is an intermediate form between granular and tubular. The photographs are grouped by the primary feature that is being illustrated such as granular forms, simple tubes, branching, distribution, and overprinting. In many photographs, more than one textural feature is present but typically only the feature being demonstrated is described. A glossary of terms is provided in Table 2 to aid in the description of the alteration features. In the text below, terms that are from the glossary are italicized.

The tunnel forms are more complex and varied than granular textures and are therefore further characterized by their shape, density, contents, and their relationship to other features in the thin sections such as fractures and vesicles. Seven major characteristics are used to describe the variety of tunnel morphologies. These are: (1) size (length and width); (2) spacing between similar tunnels; (3) curvature; (4) roughness; (5) changing width along the length of the tunnel; (6) branching; and (7) tunnel contents. In addition we recognize differences in how tunnels are distributed relative to fractures, minerals, and vesicles, and we note that some thin sections have a single type of alteration but others exhibit multiple forms.

The term thin is applied to tunnel widths (diameters) less than 3 µm from edge to edge (Fig. 8A) and long tunnels are more than 50 times longer than they are wide (Fig. 6A). The length of short tunnels is less than 10 times their width (Fig. 6B). The density term (Fig. 2) is based on how closely packed the tunnels are along the fracture or other surface from which they originate: close tunnels have a center-to-center distance that is less than 10 times the tunnel width (Fig. 7A), whereas the center-to-center distance of separated tunnels are apart by more than 10 times their diameters (Fig. 7B). Tunnels are usually curved and directed, such as away from a fracture (e.g., Figs. 5A, 6A). These directed tunnels can be nearly linear (curvilinear), kinked with sharp changes in direction (Figs. 5A, 9B), or appear have a tangled knotted appearance along their length (Fig. 8A). Convoluted tunnels turn back toward their point of origin (e.g., Fig. 24A) and do not appear to be directed away from their point of origin. Some tunnels have nearly constant width ±20% over their entire lengths (Figs. 5A, 7A), but others are variable (e.g., Figs. 14, 15, 20). Some taper from their origin at a free surface to a point in the glass (Fig. 13B). Others have repeated variations in their width resulting in a rhythmic annulated tunnel (Fig. 13A), or have a single bump between the tunnel origin and end termed engorged tunnels (Figs. 11B, 22B), or multiple irregular bumps (Fig. 12A). Rarely, a tunnel will broaden into a mushroom shape (Fig. 14A) or central disk (Fig. 12B). In addition to tunnels of variable width, there are round bud and bubble textures that are present at the margins of fractures (Figs. 16, 17A).

The surfaces of some tunnels are smooth, having irregularities that are less than 0.5 μm (Fig. 10A). Rough tunnel walls (Figs. 7B, 10B, 11) are embellished with fine (>1 μm) extensions into the glass uniformly or periodically distributed along the tunnel. The widths of individual extensions from the walls of tunnels are much thinner than the width of the tunnel. In contrast to this, branches (Fig. 2) are commonly the same width as the main tunnel. Branching may be simple with the nodes widely spaced along a tunnel, (Fig. 17B), or mossy (Fig. 4A), or networks (Figs. 18A, 19), where branches are crowded together and branch repeatedly. In the case of simple, mossy, or network branching, daughter branches are the same diameter as the parent branch. In other cases branches are narrower than their parent tunnels (Figs. 20, 21A). Some tunnels have crowns, composed of multiple radiating tunnels that have a different form than the tunnel from which they originate (Figs. 21A, 23A, 25B).

The contents of tunnels can vary as well. Some tunnels appear to be empty (Fig. 9A), but others are partially or completely filled (Fig. 21B) with opaque material. Tunnel contents can be segregated into oval or ovoid bodies typically 1–2 μm in diameter that are commonly evenly distributed along the interior of the tunnel (e.g., Figs. 22A, 25A). However, a single, large, 20 μm ovoid body is present in some tunnels and the tunnel walls swell around the body (Fig. 11). In one example the large ovoid body is divided by dark septae into five separate bodies (Fig. 22B). Septae can also divide a tunnel into multiple chambers that are 5–10 μm long (e.g., Figs. 22A, 23). Broad flat tunnels are divided by a honeycomb or patterned with a filigree of dark material (Figs 14B, 15). Tunnels may also have spiral filaments (e.g., Figs. 13A, 23B).

The distribution of alteration textures within glass has also been documented in this study. Tunnels are usually found distributed along fractures (e.g., Figs. 5A, 6A) but sometimes they are distributed around vesicles (Figs. 26, 27A, 29A), varioles (Figs. 27B, 28A), or phenocrysts and microphenocrysts (Figs. 6B, 28B). Tunnels are commonly directed away from fractures, but sometimes they are parallel to fractures (Fig. 31). Tunnels may also radiate from a point at the edge of a fracture (Fig. 5B). Some tunnels turn sharply from their initial direction perpendicular to the surface where they originate to a direction parallel to the alignment of other tunnels in the glass (Fig. 26B). In one example, the alignment of tunnels is parallel to the major axis of elongation of vesicles in the glass (not shown).

The photographs also illustrate the temporal relationship of alteration in some samples. For example, Figure 29A shows tapered tunnels radiating from a vesicle. Originally the tunnels extended into glass, but the vesicle, tunnels, and surrounding glass have now been replaced with phyllosilicate. Overprinting of a network is obvious in Figure 29B. Here a granular border evolves into a dark network in glass in the upper half of Figure 29B. In the lower half of the figure, glass containing a previous dark network has been transformed into a yellow phyllosilicate. In Figure 30A a semicircular alteration pattern, which is similar to that in Figure 8B, has been replaced by a subsequent phase of alteration, which suggests conditions changed during the formation of this alteration boundary.

Morgenstein (1969) illustrated granular and tubular textures in three dredged basalts from the Mid-Atlantic Ridge and one from a fracture zone along the Pacific-Antarctic Ridge. His transmitted-light images showed granular alteration forming a semicircle around a fracture (similar to Fig. 4A) and 20-µm-wide zones between glass and palagonite along fractures (similar to Fig. 4B). His photographs also include what he called “hair tunnels” (see Table 2) extending 50 μm from a granular texture into glass (similar to Fig. 6A) and from fractures into glass (similar to Fig. 9A). He also described a “solid solution” border that is similar to the dark zone between glass and palagonite (Fig. 4).

It was not until 1996 that tubular and granular features were again emphasized and documented with transmitted-light photomicrographs (Giovannoni et al., 1996). Also at this time and from the same Costa Rica Rift site, granular and tubular features were illustrated in backscattered electron images (Furnes et al., 1996). More detailed close-up photomicrographs (Fisk et al., 1998a) illustrated that the tubular phenomenon was more varied than the tunnels seen in Costa Rica Rift basalts. In these new photographs, mossy and branching tunnels, granular alteration, as well as tunnels with pronounced septae and cell-sized inclusions were illustrated from separate seamounts in the Pacific Ocean (one dredged and one cored basalt), from the Mid-Atlantic Ridge (two cored basalts), and from the Indian Ocean (one cored basalt) (Fisk et al., 1998a).

In the 2000s the growing literature of oceanic glass alteration reported additional morphologies including: budding along a tunnel (Furnes et al., 2000a), similar to what we have called tangled texture (Fig. 8A), and bifurcating tunnels (Furnes et al., 2001a, 2002). Annulated and convoluted tunnels were described by Banerjee and Muehlenbachs (2003) in basalts from the Ontong Java Plateau, which we have also illustrated with samples from Chagos Ridge (Fig. 13A) and from the Philippine Sea (Fig. 24A), respectively. Josef (2006) identified a number of textures not previously described, such as mushroom (Fig. 14A) and engorged (Fig. 11B). Networks of branched tunnels were first described by McLoughlin et al. (2009), and the type example is illustrated in Figure 19B. Here we also report networks made of thinner branched tunnels (Figs. 18A, 28B).

An ichnotaxonomic framework was advanced for glass alteration by McLoughlin et al. (2009) that considers the textures as trace fossils, and two new ichnogenera were proposed, corresponding to the two broad granular and tubular morphotypes discussed here and in earlier reports. Five ichnospecies were also defined on the basis of morphological characteristics and these are compared to the range of textures illustrated here, which includes new morphological variants:

1. Granulohyalichnus vulgaris isp. has a granular form (McLoughlin et al., 2009, their figure 2), and is very common and comparable to the examples illustrated here (e.g., Figs. 3 and 4).

2. Tubulohyalichnus simplus isp. has an unornamented tubular form (McLoughlin et al., 2009, their figure 3). This alteration morphology is also illustrated here (Figs. 5–31) and with further descriptors including short or long, thick or thin, closely spaced or separated, curvilinear, smooth, or constant width, possibly with tunnel contents (Fig. 2).

3. Tubulohyalichnus annularis isp. has an annulated tubular form (McLoughlin et al., 2009, their figure 4). This morphotype is also illustrated here (Fig. 13A); however, we have subdivided the annulated type further into those with regularly sized and spaced annulations (Fig. 13A), and those with irregularly sized and spaced annulations termed engorged and bumpy (Figs. 11B and 12A, respectively). The latter can be compared to the “string-of-beads” form previously described by Fisk et al. (1998a).

4. Tubulohyalichnus spiralis isp. has a helicoidal tubular form with a coiled or helical axis (McLoughlin et al., 2009, their figure 5), with up to 12 rotations reported, and can be either sinistrally or dextrally coiled. This ichnospecies may show a linear or curved growth axis with the spacing and diameter of the whorls changing along its length. This morphotype was not identified in the selection of sub-seafloor samples described here, although we did find spiral-shaped filaments within some of the tunnels (Fig. 23B) and extending from the crown in another example (Fig. 21A).

5. Tubulohyalichnus stipes isp. has a branched tubular form (McLoughlin et al., 2009, their figure 6) with dichotomously branching tubes, in which the diameters of the daughter branches are equal to that of the parent branches, and occur in dense intergrowths that may be hemispherical-shaped clusters or more irregular bands. The type material is illustrated in this atlas (Fig. 19B); however, many additional variations of branching are also illustrated here, for example, with more widely spaced branches (Fig. 17B) or palmate crowns (Fig. 20A).

The contents of tunnels have been noted in previous photomicrographs (Fisk et al., 1998a; Josef, 2006), and here tunnel contents are used as a distinguishing characteristic of the alteration. Tunnels are often transparent and apparently empty; however, others have dark contents (Fig. 21B) that we interpret to be either foreign material introduced during drilling, sampling, or thin-section preparation, or alternatively as primary organic residue (cf. Preston et al., 2011). The single or multiple ovoid bodies spaced along the tunnel (Fig. 22A) are intriguing, and further investigation, for example with fluorescent microscopy techniques, could test the hypothesis that they are cellular bodies. The origin of septae (Fig. 23), spiral filaments, ornaments inside tunnels (Fig. 23B), and petaloid or flattened tubes with ribbing and/or honeycomb ornament (Fig. 14B, 15) may be revealed with microanalytical techniques (Alt and Mata, 2000; Knowles et al., 2011a, 2011b).

We expect that certain styles of alteration of basalt glass will be correlated with environmental variables at the time of alteration. Factors such as temperature, aqueous chemistry, and oxygen abundance are known to affect the secondary mineral chemistry and therefore assumedly the textural styles of alteration. In one attempt to relate alteration styles to environmental conditions, Josef (2006) studied 140 thin sections from 63 DSDP and IODP boreholes and found no correlation. This may be because in the sub-seafloor the temperature and composition of the formation water are usually not measured and must be inferred, and the conditions at the time of sample collection may not represent the conditions when the alteration occurred.

In another study at DSDP Site 504B and ODP Site 896A of the 5.9 Ma Costa Rica Rift, Furnes and Staudigel (1999) and Staudigel et al. (2006) compared the abundance of granular and tubular alteration, which was determined by point counting, with measured and inferred subsurface conditions in the ocean crust. At these sites, measured temperatures increase with depth, and porosity is highest at ∼75 m below the sediment/basalt interface and decreases below this depth. Permeability also decreases with depth, and it is inferred that the oxygen content of formation fluids decreases with depth. They found that granular alteration was present from the basalt/sediment interface to a depth where the current temperature was ∼115 °C (500 m). Tubular alteration had a more restricted range, being rare near the basalt/sediment interface and at depths where the temperature was greater than 90 °C. The tubular style had a peak abundance of ∼10% at ∼120–130 m depth. This depth corresponds to a temperature of ∼70 °C where formation permeability is high and the presence of celadonite indicates that alteration occurred in a relatively oxidizing environment (Furnes and Staudigel, 1999; Furnes et al., 2001a). Their work shows that granular alteration is limited to temperatures less than 115 °C (500 m) and that tubular alteration is limited to shallower regions of the crust where temperatures are lower and fluid flow is likely to be higher than in deeper parts of the crust.

Boreholes on the flanks of Juan de Fuca and Mid-Atlantic Ridges that are fitted with devices (CORKs) that monitor physical and chemical conditions in the holes (Fisher et al., 2005, 2011; Expedition 336 Scientists, 2012) can also provide basalt samples in which alteration features can be documented. Although alteration textures may form at conditions different from the current conditions at the site, CORKed holes may enable studies where sub-seafloor environmental conditions are compared to the distribution, abundance, and variability in alteration features like those described in this atlas.

The period over which the alteration textures in basalt glass forms is not known. The textures may form any time between eruption into the cold, oxygenated seawater to when the basalt is deeply buried and surrounded by warm anoxic fluids. In the shallow ocean crust at both the 5.9 Ma Costa Rica Rift and the 110 Ma Western Atlantic oceanic crust, there are similar maxima in the amount of granular and tubular alteration as a percentage of the total alteration (see Furnes et al., 2001a, their figure 11). This suggests that a substantial portion of this alteration is established early in the crustal history, i.e., within ∼6 m.y. Images presented here indicate that there can be at least two episodes of texture formation at some locations (Figs. 29, 30). These episodes probably reflect changing conditions in the seafloor; however, the time between these episodes of texture formation is not known.

Alteration textures in oceanic basalt glass have been documented in a comprehensive sample set from the ocean basins (Fig. 1). The alteration textures originate at the glass/water interface both on the seafloor and along the fractures and surfaces of glass that is buried in the crust. The alteration transforms the glass to palagonite and creates cavities in the glass that exhibit a wide range of simple, intricate, or ornate forms, which are summarized in graphical form (Fig. 2) and documented in photomicrographs (Figs. 3–31). The photographs illustrate prevalent as well as rare forms. The granular form is common and it is often abundant along most of the altered fractures in some samples. Also quite common is a mossy texture. Tubular or tunnel forms are less common. Often cohorts of tunnels have a common orientation away from their region of origin and although within a few microns of each other they do not intersect. Some tunnels have random tracks and do not appear to be directed away from their point of origin. Tunnel diameters are often constant, not varying more than 20% over a 100 µm length of tunnel. In some basalts only a single type of alteration, such as the granular form, is present. However, some rocks contain multiple forms of alteration within a few tens of micrometers of each other. The superposition of alteration types in single thin sections probably reflects changing conditions in the sub-seafloor. The wide range of alteration features in some individual basalts suggests variations in chemical or physical conditions over time or on micrometer scales at a given time. The images presented here are a starting point for correlating alteration features with physical and chemical conditions in the seafloor. We consider the photographic atlas to be the current state of knowledge, which will likely be expanded as new features are discovered. Biogenicity has been proposed for alteration features that are similar to some of those illustrated here; however, if future work with microbial cultures and environmental samples documents the biogenic production of some features, we would not extend this interpretation to the broad range of features illustrated here.

Thanks to the U.S. Fulbright Program, the U.S.–Norway Fulbright Foundation, and the Centre for Geobiology at the University of Bergen for their financial support. The Centre for Geobiology was a gracious host during this study. Jeff Karson provided a scanned version of Maury Morgenstein’s Master’s thesis from the Syracuse University library. Samples were acquired from the Deep Sea Drilling Project and the Ocean Drilling Program sample repositories and funds for that travel to visit the repositories was provided by the National Science Foundation. J. Josef described many of the samples in this study in her Master’s thesis at Oregon State University (2006). Many of those samples were reexamined and re-photographed for this atlas and some of her descriptive terms were adopted.