3 Article

Geochemistry
G
Volume 4, Number 4
Geophysics 19 April 2003
1037, doi:10.1029/2002GC000470
Geosystems
AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES ISSN: 1525-2027
Published by AGU and the Geochemical Society

Tuff life: Bioalteration in volcaniclastic rocks from the

Ontong Java Plateau
Neil R. Banerjee and Karlis Muehlenbachs
Department of Earth and Atmospheric Sciences, University of Alberta, 1 – 26 Earth Sciences Building, Edmonton,
Alberta, Canada T6G 2E3 ([email protected]; [email protected])

[1] We report microscopic textural, geochemical, isotopic, and biomolecular evidence for microbial
alteration of glass shards in a 337.7 m thick sequence of poorly sorted vitric and lithic tuffs recovered during
Leg 192 of the Ocean Drilling Program on the Ontong Java Plateau. Petrographic analysis has revealed the
highest density and variety of exceptionally preserved microbial alteration textures in the glass shards, when
compared to previous studies of glassy pillow basalt margins from ocean crust and ophiolites. Two textural
types of microbial alteration are commonly observed: tubular and granular. Tubular textures are characterized
by well-preserved, micron-scale, tubular to vermicular, channel-like features with both smooth and scalloped
walls that commonly extend from a granular alteration interface rimmed by clay into unaltered glass. These
channels are highly convoluted or twisted and in some instances bifurcate. Detailed scanning electron
microscopy (SEM) images reveal the presence of delicate filaments and desiccated thin films with
morphologies suggestive of a biogenic origin within the channels. Granular textures appear as solid bands,
semicircles, or irregular patches of individual and/or coalesced spherical bodies with irregular protrusions
into fresh glass. Microprobe X-ray element maps show elevated levels of carbon, nitrogen, phosphorous, and
potassium associated with the microbial alteration features. Bulk-rock carbon isotope ratios of disseminated
carbonates in tuffs preserving fresh glass are depleted (<
9%), suggesting biologic fractionation. The
presence of nucleic acids within the microbial alteration features has been confirmed though staining with
ethidium bromide, a stain that specifically binds to double-stranded DNA and RNA. The presence of DNA/
RNA suggests that the biogenic features may be relatively recent and that microbes may be currently active.

Components: 7965 words, 14 figures, 1 table, 1 video.

Keywords: bioalteration; geomicrobiology; Ontong Java Plateau; tuff; basaltic glass; stable isotopes.
Index Terms: 3099 Marine Geology and Geophysics: General or miscellaneous; 1050 Geochemistry: Marine geochemistry
(4835, 4850).
Received 4 November 2002; Revised 4 February 2003; Accepted 7 February 2003; Published 19 April 2003.

Banerjee, N. R., and K. Muehlenbachs, Tuff life: Bioalteration in volcaniclastic rocks from the Ontong Java Plateau,
Geochem. Geophys. Geosyst., 4(4), 1037, doi:10.1029/2002GC000470, 2003.

1. Introduction of microbes in the alteration process [Thorseth et

al., 1992, 1995, 2001; Furnes et al., 1996, 1999,
[2] Recent studies of alteration of volcanic glass in 2001a, 2001b, 2001c, 2002a, 2002b; Fisk et al.,
massive and pillow basalts from the oceanic crust 1998; Torsvik et al., 1998; Furnes and Staudigel,
and ophiolites have demonstrated the importance 1999]. These studies have shown that glass alter-

Copyright 2003 by the American Geophysical Union 1 of 22

 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

ation on the seafloor is not simply an inorganic than 1.6 million km2 and with a crustal volume
process and have heightened our awareness of the of 40–50 million km3 [Mahoney, 1987; Coffin and
deep subseafloor biosphere that is found away Eldholm, 1993]. Preliminary data from Leg 192
from the isolated oases provided by mid-ocean and previous sampling of the OJP suggest the
ridge hydrothermal springs. Although glassy majority of the plateau formed during a single,
basalts are common within the upper oceanic crust, relatively rapid volcanic episode at 122 Ma, mak-
they are characteristically impermeable and have ing it possibly the largest magmatic event on the
low porosity. Volcanic tuffs, on the other hand, Earth in the last 200 m.y. [Tarduno et al., 1991;
maintain high porosity and permeability for geo- Mahoney et al., 1993, 2001]. The samples for this
logically extended periods of time until open study were recovered during Leg 192 of the Ocean
spaces are filled by the precipitation of secondary Drilling Program (ODP) from Hole 1184A located
minerals. This increases the flow of water and at a water depth of 1661.1 m on the eastern lobe of
nutrients within the tuff and enhances the like- the OJP (Figure 1). Hole 1184A was drilled ahead
lihood of microbial alteration. without coring to 134.4 meters below seafloor
(mbsf) and was rotary cored from 134.4 to 538.8
[3] In this paper we describe textural, geochemical,
mbsf with an average recovery of 81.3% [Mahoney
isotopic, and biomolecular evidence for microbial
et al., 2001]. The recovered sequence consists of
alteration of glass shards in a sequence of poorly
201.1m of calcareous early Miocene nannofossil
sorted vitric and lithic tuffs [Mahoney et al., 2001].
foraminifer ooze (Unit 1) overlying generally
These samples contain some of the best preserved,
coarse-grained volcaniclastic rocks, including tuff,
highly diverse, and abundant evidence of microbial
lapilli tuff, and lapillistone (Unit 2) and separated
alteration of volcanic glass. Glass shards in these
by an approximately 1-cm-thick ferromanganese
tuffs were likely exposed to high water-rock ratios
crust [Mahoney et al., 2001].
that could have provided nutrients to a thriving
microbial community. Although still unquantified, [6] The volcaniclastic rocks consist of ash- to
we believe marine tuffs may host a previously over- lapilli-sized lithic clasts and vitric shards, accre-
looked flourishing deep biosphere that may contrib- tionary lapilli, armored lapilli, and crystal frag-
ute significantly to global geochemical fluxes. ments (plagioclase and pyroxene) in a matrix of
fine-grained vitric and lithic ash, clay, and other
[4] At present there exists considerable contro-
alteration minerals cemented by smectite, anal-
versy over the criteria necessary for identification
cime, calcite, rare celadonite, and several zeolites
of life in the rock record [e.g., Brasier et al., 2002].
[Mahoney et al., 2001]. Well-preserved whole and
By using a detailed, high resolution, and multi-
fragmented accretionary and armored lapilli are
technique approach we are able to convincingly
common, suggesting shallowly submerged erup-
demonstrate the presence of microbial activity.
tion, and argue for an ash fall depositional setting
Continued application of these multiple techniques
with little postdepositional reworking (Figure 2)
will help elucidate the extent in time and space to
[Mahoney et al., 2001]. Twelve samples of the
which bioalteration of volcanic glass in a marine
volcaniclastic material were selected aboard ship
environment can be found. These techniques could
based on macroscopic identification of vitric clasts
also be applied to samples returned from Mars and
in order to look for microbial alteration in volcanic
other extraterrestrial bodies in the search for life
glass. Glass shards in five of these samples are
where liquid water and conditions suitable for life
completely altered to clay while the remaining
may have existed [Banerjee et al., 2002a].
seven samples contain variably altered glass shards
with fresh glass cores (Figure 3). Preliminary
2. General Geology and Samples analyses of glass shards throughout the volcani-
clastic unit suggest they are restricted to a narrow
[5] The Ontong Java Plateau (OJP) is the world’s range of basaltic compositions (R. White, personal
largest volcanic oceanic plateau covering more communication, 2002) and eruption of the volcani-

2 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

150° 155° 160° 165° 170° 175°

10° 20 No 10°

0o
1184

5° 5°
807 20oS
AUSTRALIA
803

1187 120oE 150oE

0° 289 1185 0°

1186
1183

High Plateau Eastern Salient

-5° -5°

288 1184

-10° -10°

150° 155° 160° 165° 170° 175°

-7 -6 -5 -4 -3 -2 -1 0

Predicted Bathymetry
(km)

Figure 1. Location (inset) and predicted bathymetry maps [after Smith and Sandwell, 1997] of the Ontong Java
Plateau showing the locations of sites drilled during Leg 192 (stars). The locations of other DSDP/ODP holes are
indicated by black dots. Contour interval is 1000m. Modified from Mahoney et al. [2001].

clastic units was likely penecontemporaneous with sections and grain mounts were sputter coated with
the main plateau magmatic event at 122 Ma a thin film of iridium, approximately 40 Å-thick.
(L. Chambers, personal communication, 2002).
[8] X-ray mapping on the same iridium-coated thin
sections was carried out with a JEOL JXA-8900R
3. Analytical Methods electron microprobe, using an accelerating voltage
of 15 kV, and a probe current of 1.5  10
8 A.
[7] Scanning electron microscopy (SEM) observa- Carbon and nitrogen peaks were standardized rel-
tions were performed on a JEOL JSM-6301FXV ative to silicon carbide and boron nitride, respec-
instrument connected to a Princeton Gamma Tech tively. Instrument calibration for all other elements
IMIX energy-dispersive spectrometer system. The was performed on natural standards. Carbon was
analyses were performed at an accelerating voltage routinely measured on two different spectrometers
of 20 kV and a working distance of 15mm. Thin to monitor the reproducibility of observed signals.
3 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

cm Oxygen was liberated from silicates for O-isotopic

5
analysis using the BrF5 method by reacting samples
at 600
C overnight [Clayton and Mayeda, 1963;
Muehlenbachs and Clayton, 1972]. The samples
were not acid treated prior to reaction to remove
carbonate in order to avoid removing hydroxides
and potentially some clays. The samples contain
very little carbonate so by mass their d18O should
not affect bulk rock values to any significant level.
The liberated oxygen was reacted with carbon to
produce CO2 and analyzed on a Finnegan MAT
252 mass spectrometer. The data are reported in the
10 usual delta-notation with respect to PDB for carbon
and SMOW for oxygen [Craig, 1957, 1961].
[10] Duplicate thin sections were analyzed for the
presence of nucleic acids by staining with 0.0001%
30, 80-Diamino-5-ethyl-6-phenylphenanthridinium
bromide (ethidium bromide; EtBr) for approxi-
mately 1 minute. This stain contains a fluorescent
dye that specifically binds to DNA and RNA. Excess
stain was then removed by washing three times with
filter sterilized PBS. The sample was imaged using a
15 single-photon Molecular Dynamics Multiprobe
2001 laser scanning confocal microscope (LSCM).
The excitation source was an argon-krypton laser
with excitation at 488/568 nm and emission above
590 nm. Initial observation was conducted on a
Nikon Diaphot inverted microscope. Image recon-
struction was accomplished by acquiring images at
0.5 micron intervals over a focal distance of 20 mm
followed by deconvolution with Imagespace II
scanning software on an SGI workstation.

20 4. Results
Figure 2. Example of volcaniclastic material cored 4.1. Transmitted Light Petrography
from Hole 1184A (lithic-vitric lapilli tuff; sample 192 –
1184A-43R-3, 5 – 20 cm). Note whole and broken [11] Textures of possible microbial origin were
accretionary lapilli suggesting a shallowly submerged first observed in thin sections prepared aboard
eruption setting and the absence of postdepositional ship from Hole 1184A core material that contains
reworking. From Mahoney et al. [2001, Figure F13].
fresh glass [Mahoney et al., 2001]. Subsequent
detailed petrographic analysis has revealed an
[9] Stable C- and O-isotope analyses of carbonates astounding density and variety of alteration tex-
were performed by pouring 100% phosphoric acid tures attributed to microbial activity [Banerjee et
on whole rock powders under vacuum [McCrea, al., 2002b; Banerjee and Muehlenbachs, 2002].
1950] and analyzing the exsolved CO2 on a Finne- Two textural types of microbial alteration are
gan MAT 252 mass spectrometer. Yields of CO2 in commonly observed at the glass alteration inter-
the samples varied from 0.03 to 6% by weight. face: tubular and granular [e.g., Furnes and Stau-
4 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

5 mm

Lithic Fragment

Glass

Figure 3. Whole thin section photo showing typical texture of vitric-lithic tuffs (sample 192– 1184A-31R7, 43– 47
cm). Individual glass grains are commonly rimmed by clay minerals and microbial textures extend from this
boundary into fresh glass.

digel, 1999]. Tubular textures are the most striking [12] Individual channels within tubular textures
textural type and are characterized by micron- range in diameter from 1 to >10 mm and may attain
scale, tubular to vermicular, channel-like features lengths >100 mm. Both straight and curved chan-
and branching bodies extending into fresh glass nels are observed. Highly convoluted channels are
from the alteration boundary (Figure 4). Granular not uncommon and in some cases the channels
textures appear as solid bands, semicircles, or bifurcate or contain numerous branches. Where
irregular patches of individual and/or coalesced bifurcating or branching channels occur, the
spherical bodies with irregular protrusions into branches are always of the same approximate
fresh glass (Figures 4b and 4c). Individual spher- diameter as the original channel. Some channels
ical bodies or patches within granular textures are appear open and occasionally contain dark spher-
commonly 0.2–3 mm in diameter. These tubular ical bodies (Figure 4d). Although smooth walls are
and granular textures are of similar size and most common, channels with scalloped walls are
morphology to granular and tubular features attrib- observed with varying degrees of segmentation
uted to microbial alteration in previous studies (Figures 4e and 4f). Channels are also observed
[e.g., Furnes and Staudigel, 1999; Furnes et al., to terminate or link clusters of spherical bodies,
2001b]. These two textural types commonly occur which are referred to as string-of-pearls texture
together with tubular structures of various sizes [Mahoney et al., 2001, Figure F68]. Most channels
and morphologies extending from the granular appear dark in thin section possibly due to staining
alteration boundary (Figure 4c). or because they are filled.
5 of 22
Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

50 um
C 50 um

F
50 um

25 um
t

gr
gr

E
0.5 mm

25 um

D
A

6 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

[13] The glass-clay alteration boundary is com- method requires a minimum amount of processing
monly marked by a 50 to 100 mm wide zone of and appears to have preserved delicate structures
irregularly shaped, micron-sized features that give that are obliterated by the polishing process (Fig-
the boundary a granular appearance (Figures 4b ures 5–9). As observed in thin section, the chan-
and 4c). Most channels originate from this boun- nels are concentrated at the glass-clay alteration
dary and extend into the fresh glass. Occasionally boundary. Within the channels, filaments and small
individual channels or tubes are observed isolated rounded structures or aggregates are commonly
within the glass matrix with no apparent connec- observed (Figures 5–8). The filaments are com-
tion to this alteration boundary. Extensive exami- monly visibly attached to the channel walls and
nation of this texture has shown the features occur both as filled and hollow varieties (Figures
intersect the top and/or bottom plane of the thin 5 – 8). The hollow filaments commonly appear
section and that the isolated appearance of these segmented (Figure 7b). In an exceptional case, an
features is an artefact of the thin section prepara- extremely thin film was observed to stretch
tion (see Movie 1, available in the HTML version between two filaments (Figure 6d). This type of
at http://www.agu.org/journals/gc). This is also true structure, interpreted as a preserved biofilm, is
of granular alteration features, which sometimes more commonly observed as desiccated remains
appear as isolated spherical bodies. These textures on the channel walls, usually at the base of
represent complicated aggregates of features com- filaments (Figures 5 and 6). In addition to more
monly connected by thin tubes or which are common micron-scale filaments, nannoscale struc-
attached to surrounding features. tures of possible microbial origin are also observed
(Figure 9). These have similar morphological char-
acteristics as larger structures in channels except
4.2. SEM Imaging for their extremely small size. Qualitative EDS
[14] We have conducted detailed SEM imaging of analysis of the material within the channels
possible biogenic features in thin sections and on (including the filaments, thin films, and spherical
grain mounts of freshly exposed glass grains using bodies) indicates elevated potassium concentra-
iridium as the coating material (Figures 5 – 9). tions and a composition resembling clay minerals.
Iridium provides an exceptionally thin (40Å), These structures do not resemble diagenetic clay
stable, and conductive film [Blake et al., 1999] minerals and do not appear to have been formed
that allows intricate structures to be resolved, during inorganic precipitation of clay during alter-
particularly within the channels. As a result, we ation of the glass. It is unlikely these structures
have imaged delicate filament-like structures, pos- represent active or recent microbial remains but are
sible cellular structures, and material resembling likely fossil evidence for the presence of microbes.
desiccated biofilm that might otherwise be We suggest that microbial activity is responsible
obscured by thicker coatings. for these structures and is in effect enhancing the
alteration of glass to clay.
[15] Glass shards separated from the tuff by hand
plucking from rock samples produced the most [16] In several cases, channels observed by trans-
spectacular images. This sample preparation mitted light microscopy that intersect the upper

Figure 4. (opposite) Transmitted light photomicrographs of glass grains displaying microbial alteration textures. (a)
Glass grain rimmed by red-brown clay in volcaniclastic matrix containing altered glass and lithic fragments cemented
by clay and zeolites. (b) Granular (gr) and tubular (t) alteration textures extending from the clay alteration boundary
into fresh glass. Note the tubular features terminate in a cluster of spherical bodies. (c) Tubular structures of various
sizes and morphologies extending from a granular alteration boundary. Note segmented appearance of some channels.
(d) Open channel with dark spherical bodies (possibly fossil microbes) lining the walls. (e) Anastomosing channels
with dark walls and segmented appearance. See X-ray element mapping section for C, N, P, and K pattern (Figure
10). (f) Various open (clear) and filled (dark) channels of different sizes as well as clusters of spherical bodies. Highly
segmented channels with darkly stained walls are visible in the center. Individual septae are visible in some channels.
7 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

500 um 20 um
A B

5 um 1 um
C D

Figure 5. (a to d) SEM image series at increasing magnification showing development of channels containing
numerous filaments and cell-like structures (arrow in d) at the glass-clay alteration boundary. Fresh glass appears
smooth (dark areas) whereas clay has a mottled appearance (bright areas). Note the smooth, pebbled surface texture
possibly resulting from a preserved desiccated biofilm observed at high magnification (d).

plane of the thin section were also identified by highest anomaly with lesser amounts of N, P and K
SEM (see Figure 10). Individual channels are relative to fresh glass away from bioalteration
highly irregular in shape, commonly displaying features. Although the intensity of the C signal
curved or scalloped edges and in some cases does vary, elevated levels of C are observed in all
appear to be segmented. In thin sections, the cases where bioalteration features occur, whereas
delicate structures observed in channels from N, P, and K highs may or may not be observed.
plucked glass shards are absent, likely due to Element maps for Ca, Mg, Al, Na, Si, Cl, and Cu
abrasion during preparation. or Ti were also routinely measured. Most of these
elements do not show enrichments and argue
4.3. Element Mapping against the possibility of carbon highs due to
[17] X-ray element maps collected by electron inorganic carbonate material (e.g., Ca, Mg, Fe) or
microprobe on thin sections and grain mounts epoxy (e.g., Cl). Small aluminum highs are occa-
commonly show elevated levels of C, N, P, and sionally observed due to build-up of the polishing
K associated with the microbial alteration features compound used on thin sections. Sulfur, iron, and
along the glass-clay alteration front (Figures 10– rare copper highs occur at the margin of some
12). The observed enrichments are highly restricted channels and granular alteration features and are
to areas of presumed microbial attack and diminish associated with C, N, P, and K highs. These
suddenly away from these areas. Carbon shows the anomalous values are related to the presence of
8 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

Figure 6. (a to d) SEM image series at increasing magnification showing the presence of filaments attached to the
channel walls. Note the presence of a thin film stretched between two filaments (arrow in d) interpreted as a preserved
biofilm. The irregular surface observed at the base of the filaments may also represent a preserved desiccated biofilm.

micron-sized grains of iron sulfides next to chan- either d13C or d18O of disseminated carbonate.
nels and within granular alteration zones (Figure Two samples with carbonate contents significantly
12). Sulfide grains are only found in association higher than the rest (6.39 and 5.69 wt.%) have the
with microbial alteration textures and are not lowest and highest carbon isotope ratios (
15.8
observed isolated in fresh glass. Titanium is occa- and
3.9), respectively. The distribution of carbon
sionally associated with channel features. Titanium isotope ratios does not show a bias between glassy
can be passively accumulated as a residual element and nonglassy samples (Figure 13). We believe this
during abiotic glass alteration [e.g., Staudigel and is because nonglassy samples were altered under
Hart, 1983] and is possibly also enriched during the same conditions (including bioalteration) as the
microbially mediated alteration. glassy samples that preserve textural evidence of
microbial activity. If this is true, the fact that these
4.4. Stable Isotopes samples contain very little carbonate (generally less
[18] Carbon and oxygen isotopic analyses of dis- than 0.3 wt.%) could indicate a microbial source
seminated carbonate and silicate oxygen isotope for the carbonate. The d18O of silicate material
analyses in glassy and nonglassy samples are listed within the tuffs varies between 8.9% and 21.2%
in Table 1. The d13C of disseminated carbonate (Table 1). Pervasively altered samples have the
ranges from
15.8 to
3.9% (Figure 13). No highest d18O ratios, whereas samples preserving
trend with depth or correlation between isotope glass typically have lower d18O ratios. This trend is
ratio and carbonate abundance is observed for likely related to the extent of low temperature
9 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

500 nm

D
A
5 um

500 nm

C
C&D

500 nm

B
B

10 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

1 um 5 um
A B

Figure 8. (a) Delicate filament structure with several small openings indicating it is hollow. Both filled and hollow
filaments have been observed. (b) SEM image showing filament within channel and clay filled tunnels of area
mapped in Figure 11.

alteration, which results in the formation of clay ing of the stain is another problem that needs to be
minerals (i.e., smectite) with d18O ratios near 27% addressed, particularly in samples containing clay
[e.g., Muehlenbachs and Clayton, 1972; Bohlke et minerals. We performed rigorous tests on the
al., 1984]. samples to identify nonspecific binding of the
stain. Our results indicate nonspecific binding is
4.5. Nucleic Acid Staining minimal with no fluorescence observed in the thick
[19] It has been suggested the tubular and granular abiotic clay alteration band surrounding the glass
textures were formed as the result of microbial grain. The results from nucleic acid staining on thin
activity. Independent evidence of this process sections confirm the presence of DNA/RNA within
would be the presence of DNA/RNA confined to the microbial alteration features (Figure 14).
these zones of alteration. One common problem in LSCM images show fluorescence of the EtBr stain
studies that use fluorescent nucleic acid stains on away from the thick clay alteration rim around
geologic material is differentiating between fluo- individual glass grains, demonstrating the presence
rescence from specific binding of the stain to DNA/ of nucleic acids. Both differential interference
RNA and autofluorescence of mineral particles. contrast and LSCM images were acquired for
Samples prepared for nucleic acid staining and identical fields of view in order to allow associa-
controls were observed prior to staining under the tion of fluorescent areas with any visible structures
same conditions as stained samples to monitor (Figures 14a and 14b). When the differential inter-
autofluorescence. Surprisingly little evidence of ference contrast and LSCM images are overlain,
autofluorescence was observed other than a few the association between brightly fluorescent areas
instances at the edges of thin sections and individ- and channels is observed (Figure 14c). Processing
ual grains, which were avoided. Nonspecific bind- of stacked LSCM images provides a composite

Figure 7. (opposite) (a) SEM photomontage of complex channel containing a variety of filaments and cell-like
structures. The locations of images b through d are indicated by the yellow boxes in a. Note that one of the filaments
is hollow and is possibly the cast of a microbial tunnel (b). The channel contains cell-like features that may be
fossilized microbes (c) and (d). Note that these features appear to be attached to the channel walls. Images (c) and (d)
are of the same area but were taken from different angles.
11 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

Figure 9. (a and b) SEM images showing location of nannoscale filaments within channels. Arrow in (a) shows
location of (b). (c and d) SEM images showing location of intricate, lacy, hollow nano-filaments within clay-lined
channel. Arrow in (c) shows location of (d).

representation of fluorescent areas within the thin microbes introduced during drilling or sample
section and shows the fluorescence is localized preparation. Instead we feel the textures must
along channel walls (Figure 14d). We interpret this represent fossil relicts of past microbial activity.
as specific binding of EtBr to nucleic acids within Unfortunately it is not possible to determine how
cellular material left behind by microbial activity far in the past the microbes were active. It is
along the channel walls. possible the elevated carbon values and the pres-
ence of nucleic acids in the channels could have
5. Discussion been caused by contamination but we also believe
this is unlikely. Smith et al. [2000] evaluated the
[20] Textural evidence, element distributions, car- potential for contamination during seafloor drilling
bon isotopes, and the presence of nucleic acids in and found that contamination could not be com-
alteration features all suggest the involvement of pletely controlled with current drilling technology.
microbes during alteration of the OJP tuffs. Despite this fact, it would be very difficult for
Although we consider this evidence very convinc- organic contaminants to reach the alteration front
ing, it is important to consider the possibility that where the best evidence for microbial alteration has
organic matter may have been introduced to these been observed (Figure 4). The samples in this
samples by contamination either during drilling or study come from the center portion of the core
sample preparation. It is highly unlikely the tex- where fluids carrying contaminants are unlikely to
tures observed could have been formed by penetrate, except along fractures. Furthermore, the
12 of 22
 Carbon Nitrogen
A B C
Geophysics
Geosystems
Geochemistry

G
3

10 um 10 um 10 um

Phosphorous Potassium Calcium

D E F
banerjee and muehlenbachs: tuff life

10 um 10 um 10 um

Figure 10. Microprobe backscattered electron image (a) and X-ray element maps (b to f) showing the distribution of
carbon, nitrogen, phosphorus, potassium, and calcium associated with tubular features of possible microbial origin observed
in Figure 4e. Note the channels only intersect the upper plane of the thin section slightly. Also note the scalloped appearance
of the channel walls in (a). Calcium is low in the channels indicating the absence of calcite. Increasing order of elemental
10.1029/2002GC000470

13 of 22
abundance black - blue - green - yellow - red - pink. Sample 192 – 1184A-13R3 145– 148 cm.
Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

10 um

10 um
Nitrogen

Calcium
C

Potassium
10 um
10 um
Carbon

E
B

Phosphorous
10 um
10 um
A

14 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

glass shards are surrounded by thick clay alteration nutrients (e.g., organic carbon, etc.). These two
rims that would make it difficult, and probably factors may explain the higher proportion of bio-
impossible for fluids carrying contaminants to alteration features relative to glassy basalts in
reach the alteration front. In the following sections previous studies.
we highlight the evidence for microbial alteration
[22] Although an exact mechanism for the micro-
of the OJP tuffs and compare our results with
bial etching of the glass has not been determined,
previous studies of oceanic crust and ophiolites.
Thorseth et al. [1992] first proposed that microbial
5.1. Alteration Textures control over local changes in pH are responsible.
This seems quite probable since microbes are
[21] Similar textures to those described in this
known to locally modify pH [e.g., Golubic, 1973;
study have been attributed to microbial alteration
Krumbein et al., 1991] and experimental studies
of glassy pillow rims in modern oceanic crust and
have shown glass dissolution by microbes can
ophiolites [Thorseth et al., 1992, 1995, 2001;
occur [e.g., Staudigel et al., 1995]. Glass dissolu-
Furnes et al., 1996, 1999, 2001a, 2001b, 2001c,
tion may also occur due to microbially enhanced
2002a, 2002b; Fisk et al., 1998; Torsvik et al.,
extraction of bioessential trace elements by metal-
1998; Furnes and Staudigel, 1999]. Furthermore,
specific high affinity ligands [Brantley et al.,
the sizes of both types of microbial textures are
2001]. For example, Brantley et al. [2001] have
within the size range of common microbes and we
shown that trace metals in hornblende are prefer-
have no sensible abiotic explanation for these
entially released to solution in the presence of
features. These points lead us to suspect a biogenic
bacteria. Bacteria commonly use trace metals such
origin for the tubular and granular alteration tex-
as Fe, Mn, Cu, Zn, V, Mo, Ni, and Co in enzymes,
tures observed in the OJP vitric tuffs. What is
coenzymes, and cofactors (see review by Brantley
different is the density and diversity of alteration
et al. [2001]). This is a possible explanation for the
textures, particularly tubular forms, which are far
association of Cu in micron-sized iron sulfide
less common in previous studies of pillow basalts.
grains with microbial alteration textures. Copper
Microbial alteration textures are commonly diffi-
highs are relatively uncommon, so it is unlikely Cu
cult to find in glassy pillow basalts, requiring
is the driving force behind the formation of the
observation of many samples in order to find
majority of microbial textures. Instead, we suggest
commonly sparse evidence. However, we were
iron is a more likely candidate because of its
able to find these textures in a very small sample
relatively high concentration in the glass, low
suite (only 7 glassy samples from more than 300m
concentration in ocean bottom water, and the fact
of core) and microbial textures are observed in all
that microbes are likely able to extract Fe from
of these samples. In sample 192–1184A-13R-3,
glass [Brantley et al., 2001].
145–148 cm every glass shard contains microbial
alteration textures and in most shards these tex- [23] Glass dissolution by the microbes also pro-
tures likely represent more than 50% of the alter- vides an explanation for the observed septae and
ation. Glass shards in the tuffs from the OJP were segmented appearance of some channels. Individ-
likely exposed to much higher water-rock ratios uals within a colony of microbes residing next to
than the glassy rims on basaltic pillow lavas. Also, each other in a channel could cause very localized
the tuffs are overlain by more than 200m of dissolution of the glass resulting in the formation
calcareous sediment that may be a source of of these textures (e.g., Figures 4e and 4f). The

Figure 11. (opposite) Microprobe backscattered electron image (a) and X-ray element maps (b to f) showing the
distribution of carbon, nitrogen, phosphorus, potassium, and calcium associated with channel observed in Figure 8b.
These images are from a glass chip preserving filaments within a channel (see Figure 8b). Carbon highs are observed
at the edge of the channel and in the tunnels in the bottom right quadrant of (b). Potassium highs in the tunnels
located at the bottom right are associated with filling by clay minerals. This sample was not polished. Increasing order
of elemental abundance black - blue - green - yellow - red - pink. Sample 192 – 1184A-13R3 145 – 148 cm.
15 of 22
 A B C
Geophysics
Geosystems
Geochemistry

G
3

Carbon Phosphorous
10 um 10 um 10 um

D E F
banerjee and muehlenbachs: tuff life

Nitrogen Potassium Sulfur

10 um 10 um 10 um

Figure 12. Microprobe backscattered electron image (a) and X-ray element maps (b to f) showing the distribution of
carbon, nitrogen, phosphorus, potassium, and calcium associated with granular alteration features. Note the association of
micron-sized sulfide grains (bright spots in f) in a granular alteration zone with carbon, nitrogen, phosphorous, and minor
potassium. Increasing order of elemental abundance black - blue - green - yellow - red - pink. Sample 192 –1184A-39R7 99–
10.1029/2002GC000470

16 of 22
103 cm.
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

Table 1. Carbon and Oxygen Isotope Ratios, Carbonate Content, and Calculated Temperatures of Hole 1184A
Samplesa
Sample Depth, mbsf Type d13Ccarb d18Ocarb d18Osil Carb, wt.% T,
C

192-1184A, 9R-1, 106 – 108 cm 202.16 nonvitric
3.9 28.0 17.4 5.69 24
192-1184A, 11R-2, 110 – 114 cm 209.00 nonvitric
15.8 22.2 21.2 6.39 56
192-1184A, 12R-2, 133 – 136 cm 213.73 nonvitric
14.4 20.0 15.6 0.03 72
192-1184A, 13R-3, 145 – 148 cm 224.52 vitric
12.1 20.2 16.3 0.05 70
192-1184A, 31R-7, 43 – 47 cm 392.75 vitric
9.5 27.0 16.2 0.08 29
192-1184A, 39R-7, 99 – 103 cm 471.66 vitric
11.3 24.4 13.6 0.15 43
192-1184A, 40R-5, 64 – 68 cm 477.94 vitric
10.3 26.2 13.7 0.08 34
192-1184A, 41R-7, 112 – 116 cm 490.23 vitric
9.8 24.1 12.8 0.07 45
192-1184A, 42R-CC, 6 – 9 cm 500.48 nonvitric
9.4 26.1 10.9 0.29 34
192-1184A, 43R-3, 2 – 8 cm 502.75 nonvitric
8.4 21.7 15.1 0.25 60
192-1184A, 44R-3, 130 – 133 cm 514.05 vitric
11.5 25.8 8.9 0.09 36
192-1184A, 45R-1, 113 – 115 cm 520.63 vitric
12.6 23.4 13.7 0.25 49
a
Carb, carbonate; sil, silicate; mbsf, meters below seafloor.

smooth-walled channels perhaps result from con- community (e.g., Figure 4). Regardless of the
tinual dissolution of glass at the channel tip without physical framework of the colony, the diversity
prolonged etching at one location. Furnes et al. of sizes and shapes of alteration textures suggests a
[2001c] have suggested the high density and diver- complex community of microbes, likely with dif-
sity of microbial alteration textures in one sample ferent metabolic requirements and pathways.
from the Troodos Ophiolite is evidence of coloni-
zation in the form of a microbial mat. This may 5.2. Element Distributions
also be possible for the OJP tuffs, which host an [24] The common association of C, and to a lesser
abundance of microbial alteration textures of vary- extent N, with suspected microbial alteration tex-
ing morphologies and sizes, which is reminiscent tures and not elsewhere argues favourably for an
of what might be expected in a microbial mat organic origin due to the low abundance of these

1184 Glassy
1184 Non Glassy
35 504B Glass
-5
504B Crystalline

30 15
Temperature oC
δ18O Carbonate

25 40

20 70

15 115

10 180
-20 -15 -10 -5 0 5
δ13C Carbonate

Figure 13. Comparison of d13C and d18O in glassy and nonglassy tuffs (diamonds) with glassy and crystalline
pillow basalts from Hole 504B (circles). Indicated temperatures were calculated from oxygen isotope ratios of
carbonate assuming equilibrium with 0% seawater using the equations of O’Neil et al. [1969].
17 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

A B

50 um 50 um

C D

50 um 10 um

Figure 14. (a) LSCM image showing fluorescence of EtBr locally bound to cellular material within sample. (b)
Differential interference contrast image showing presence of channels (same area as a). (c) Composite image (a and b)
showing association of EtBr with channels. (d) Stacked LSCM image showing specific binding of EtBr stain to
channel walls.

elements in igneous rocks. These elements are microbes would extract K from the glass matrix.
most likely extracted from seawater by microbes Instead, K is found associated with filaments and
and become concentrated in cells. As the microbes structures of suspected biological origin within
dissolve the glass, multiply, and die, organic channels. Although proper mineral identification
remains containing C and N are left behind within of individual structures is difficult within the chan-
the alteration textures produced, resulting in the nels at these small scales, qualitative EDS analyses
elevated signals observed. Potassium and phospho- suggest the K is sequestered within clay minerals.
rous, conversely, are present in the glass matrix. The formation of clay alteration minerals mediated
Potassium in the glass is found in very low con- by microbes likely follows a complex path involv-
centrations relative to seawater so it is unlikely the ing mixtures of multiple authigenic phases and
18 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

altered amorphous glass [Alt and Mata, 2000]. much harder to explain through purely inorganic
Similarly, the concentration of P in the glass is processes. Fractionation of carbon isotopes during
also low; however, P may be less abundant in oxidation of dissolved organic matter by bacteria
ocean bottom waters. It is uncertain if the microbes has been previously used to explain depleted d13C
are able to extract P from the glass during dis- ratios in glass suspected to have been biogenically
solution but element maps show elevated concen- altered [e.g., Furnes et al., 1999, 2001a, 2001b].
trations of P in microbial alteration textures, likely We suggest the depleted d13C ratios of dissemi-
due to incorporation in cells. nated carbonate from the OJP tuffs have also
resulted from microbial activity.
[25] As mentioned above, the metabolic require-
ments of the microbes responsible for glass alter- [27] Microbial fractionation of the carbon isotopes
ation have not been determined. The micron-sized is also supported by temperature estimates for the
sulfide grains associated with both granular and formation of the carbonate, and hence temperature
tubular microbial alteration features may have of microbial growth. Downhole temperature at
formed from reduction of sulfate by microbes, Hole 1184A was not measured during Leg 192;
possibly as an energy source. This is supported however, temperatures calculated from oxygen
by the lack of sulfide grains randomly distributed isotope ratios of carbonate, assuming equilibrium
within the glass matrix as would be expected for with seawater (0%), suggest formation between
crystallization of igneous sulfide grains. Sulfate approximately 24
and 79
C [O’Neil et al., 1969].
reduction as a metabolic pathway is also consis- These temperatures are further supported by the
tent with the depleted d13C values for dissemi- low-temperature alteration assemblage in the tuffs
nated carbonate [Preuß et al., 1989] (see below). [Mahoney et al., 2001]. This range of temperatures
While sulfate reducing bacteria (SRB) may have is well within the temperature range for microbial
been part of the microbial community responsible life [Stetter et al., 1990].
for the alteration of the glass, we are not attempt-
ing to characterize the microbial alteration process 5.4. Cellular Material
as solely the result of the activity of SRB. [28] Samples treated with EtBr, that specifically
Instead, it is likely the glass contains elements binds to DNA and RNA, have confirmed the
which act as nutrients, such as iron, that are presence of cellular material in the channels.
sought by the microbes and are released during LSCM images indicate the cellular material is most
glass dissolution. We suggest that microbial alter- commonly distributed along the channel walls,
ation of glass on and beneath the seafloor, like possibly in preserved biofilms left by biogenic
many microbial communities, may occur through activity. These biofilms have very distinct mor-
the complex interplay of a consortium of phologies that have been imaged by SEM (e.g.,
microbes with varying metabolic requirements Figure 6d) and that do not appear to have textures
and survival strategies. typical of inorganically precipitated clay minerals
that might cause nonspecific binding of the EtBr.
5.3. Carbon and Oxygen Isotopes Bright fluorescent spherical and rod-shaped bodies
[26] We interpret the d13C ratios of disseminated are observed in granular alteration areas and some
carbonate in the tuffs as significantly lower (aver- channels but it is difficult to single out individual
age value of
10.5%) than what might be microbes. We believe this is positive evidence of
expected for marine carbonate (0%) or fresh basalt microbial activity because of the localized nature
(
5 to
7%) [Hoefs, 1997]. Only one sample falls of the observed fluorescence. If the fluorescence
within this range of values (0 to
7%) and it has a was the result of contamination during or after
relatively high carbonate content (5.69 wt.%). It is preparation of the thin sections, we would expect
likely the d13C ratio of this sample reflects some to observe it over the entire surface of the sample.
combination of marine and magmatic values. The It is also unlikely the fluorescence represents trap-
other samples, with much lower d13C ratios, are ping of the stain within channels because of its
19 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

localization along the channel walls. Since nucleic ingly important to document in detail more modern
acids are not stable over geological lengths of time, examples of fossil microbial life. The detailed, high
we believe the biogenic features may be relatively resolution, and multitechnique approach used in
recent and the possibility exists that the microbes this study provides a framework from which to
may be currently active. build further studies of bioalteration in geologic
materials through time and space. Extension of
6. Conclusions these techniques to the identification of traces of
early life on earth in ancient rocks may prove
[29] Glass shards within tuffs from the OJP con- promising. Perhaps the ultimate goal will be appli-
tain alteration textures indicative of microbial cation of these and complimentary techniques to
attack. The density and variety of microbial alter- astrobiology through analysis of samples returned
ation features in tuffs from the OJP are much from Mars and other extraterrestrial bodies where
higher than in similar studies of bioalteration in life may have once existed.
glassy pillow basalts from the oceanic crust and
ophiolites. While it is difficult to quantify, the
abundance of C associated with biogenic textures
Acknowledgments
in the OJP samples is high relative to previous [32] We thank H. Furnes for many stimulating discussions
studies of glass bioalteration, particularly when about bioalteration; S. Matveev for help with the microprobe;
compared to studies in ophiolites. This may indi- O. Levner for help with stable isotope analyses; G. Braybrook
for help with the SEM; R. Bhatnagar and J. Scott for help with
cate that microbial activity in OJP tuffs is rela-
the confocal microscope; S. Laurence for drafting many of the
tively recent since these element signals have not figures; and the crew, technicians, and science party of ODP
had time to degrade. This possibility is further Leg 192 aboard the drillship JOIDES Resolution. Constructive
supported by the presence nucleic acids within comments by H. Staudigel, an anonymous reviewer, and the
bioalteration features, since these molecules have editor W. White helped improve the final version. Financial
support for N.R.B. to sail on ODP Leg 192 was provided by
likely not survived since the initial eruption of the CanadaODP. This work was supported by grants from the
tuffs. The exact mechanism of microbial attack National Sciences and Engineering Research Council of
still remains unknown but pH and metal-specific Canada.
high affinity ligands may have played a role in
glass dissolution associated with microbial uptake References
of trace nutrients.
Alt, J. C., and P. Mata, On the role of microbes in the alteration
[30] Compositionally, the glass shards are similar of submarine basaltic glass: A TEM study, Earth Planet. Sci.
to mid-ocean ridge basalts; however, the water- Lett., 181, 301 – 313, 2000.
rock ratios experienced by the OJP tuffs were Banerjee, N., and K. Muehlenbachs, Bioalteration in a subsea-
floor tuff—The link between rocks and sediments, Geochim.
likely much higher than pillow basalts on the
Cosmochim. Acta, 66(Suppl. 1), 48, 2002.
seafloor. This suggests that porosity and perme- Banerjee, N., K. Muehlenbachs, H. Furnes, V. Cameron, and
ability, which control the flow of seawater and J. French, Detection of textural, geochemical, and molecu-
nutrients, may be limiting factors in microbial lar biomarkers in volcanic rocks as a proxy for studies of
attack on glassy oceanic rocks. As the glass is extraterrestrial material, Geol. Soc. Am. Abstr. Programs,
33, 175, 2002a.
dissolved, the potential for enhanced geochemical Banerjee, N. R., K. Muehlenbachs, and W. Bach, Microbial
exchange with surrounding seawater increases. We alteration of volcanic glass in a subseafloor tuff from the
believe that a thriving deep biosphere hosted in Ontong Java Plateau, Geol. Assoc. Can. Abstr. Program,
marine tuffs may contribute significantly to global 27, 6, 2002b.
Blake, D. F., M. Schulte, T. Bunch, and S. L. Cady, Submicro-
geochemical fluxes between the lithosphere and scopic morphological features and mineralogies as biomar-
hydrosphere. kers, Geol. Soc. Am. Abstr. Programs, 37, 378, 1999.
Bohlke, J. K., J. C. Alt, and K. Muehlenbachs, Oxygen iso-
[31] In light of the recent controversy over detec- tope-water relations in altered deep-sea basalts: Low tem-
tion of microbial remains in the ancient rock record perature mineralogical controls, Can. J. Earth Sci., 21,
[e.g., Brasier et al., 2002], it is becoming increas- 67 – 77, 1984.

20 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

Brantley, S. L., L. Liermann, M. Bau, and S. Wu, Up- Furnes, H., K. Muehlenbachs, T. Torsvik, O. Tumyr, and
take of trace metals and rare earth elements from horn- L. Shi, Bio-signatures in metabasaltic glass of a Caledonian
blende by a soil bacterium, Geomicrobiol. J., 18, 37 – 61, ophiolite, West Norway, Geol. Mag., 139(6), 601 – 608,
2001. 2002b.
Brasier, M. D., O. R. Green, A. P. Jephcoat, A. K. Kleppe, M. J. Golubic, S., The relationship between blue-green algae and
Van Kranendonk, J. F. Lindsay, A. Steele, and N. V. Grass- carbonate deposits, in The Biology of Blue-Green Algae,
ineau, Questioning the evidence for Earth’s oldest fossils, edited by N. G. Carr and B. A. Whitton, pp. 434 – 472,
Nature, 416, 76 – 81, 2002. Blackwell Sci., Malden, Mass., 1973.
Clayton, R. N., and T. K. Mayeda, The use of bromine penta- Hoefs, J., Stable Isotope Geochemistry, 201 pp., Springer-Ver-
fluoride in the extraction of oxygen from oxides and silicates lag, New York, 1997.
for isotope analysis, Geochim. Cosmochim. Acta, 27, 43 – 52, Krumbein, W. E., C. E. Urzi, and C. Gehrmann, Biocorrosion
1963. and biodeterioration of antique and medieval glass, Geomi-
Coffin, M. F., and O. Eldholm, Scratching the surface: Esti- crobiol. J., 9, 139 – 160, 1991.
mating dimensions of large igneous provinces, Geology, 21, Mahoney, J. J., An isotopic survey of Pacific oceanic plateaus:
515 – 518, 1993. Implications for their nature and origin, in Seamounts, Is-
Craig, H., Isotopic standards for carbon and oxygen and lands, and Atolls, Geophys. Monogr. Ser., vol. 43, edited by
correlation factors for mass-spectrometric analysis of car- B. H. Keating et al., pp. 207 – 220, AGU, Washington, D. C.,
bon dioxide, Geochim. Cosmochim. Acta, 12, 133 – 149, 1987.
1957. Mahoney, J. J., M. Storey, R. A. Duncan, K. J. Spencer,
Craig, H., Standards for reporting concentrations of deuter- and M. S. Pringle, Geochemistry and age of the Ontong
ium and 18O in natural waters, Science, 133, 1702 – 1703, Java Plateau, in The Mesozoic Pacific: Geology, Tectonics,
1961. and Volcanism, Geophys. Monogr. Ser., vol. 77, edited by
Fisk, M. R., S. J. Giovannoni, and I. H. Thorseth, Alteration of M. S. Pringle et al., pp. 233 – 262, AGU, Washington,
oceanic volcanic glass: Textural evidence of microbial activ- D. C., 1993.
ity, Science, 281, 978 – 979, 1998. Mahoney, J. J., et al., Proceedings of the Ocean Drilling Pro-
Furnes, H., and H. Staudigel, Biological mediation in ocean gram, Initial Reports [online], vol. 192, Ocean Drill. Pro-
crust alteration: How deep is the deep biosphere?, Earth gram, College Station, Tex., 2001. (Available at http://
Planet. Sci. Lett., 166, 97 – 103, 1999. www-odp.tamu.edu/publications/192_IR/192ir.htm)
Furnes, H., I. H. Thorseth, O. Tumyr, T. Torsvik, and M. R. McCrea, J. M., On the isotope chemistry of carbonates and a
Fisk, Microbial activity in the alteration of glass from pillow paleotemperature scale, J. Chem. Phys., 18, 849 – 857,
lavas from Hole 896A, in Proceedings of the Ocean Drilling 1950.
Program, Scientific Results, vol. 148, edited by J. C. Alt et Muehlenbachs, K., and R. N. Clayton, Oxygen isotope studies
al., pp.191 – 206, Ocean Drill. Program, College Station, of fresh and weathered basalts, Can. J. Earth Sci., 9, 172 –
Tex., 1996. 184, 1972.
Furnes, H., K. Muehlenbachs, O. Tumyr, T. Torsvik, and I. H. O’Neil, J. R., R. N. Clayton, and T. K. Mayeda, Oxygen iso-
Thorseth, Depth of active bio-alteration in the ocean crust: tope fractionation in divalent metal carbonates, J. Chem.
Costa Rica Rift (Hole 504B), Terra Nova, 11, 228 – 233, Phys., 51, 5547 – 5558, 1969.
1999. Preuß, A., R. Schauder, G. Fuchs, and W. Stichler, Carbon
Furnes, H., K. Muehlenbachs, T. Torsvik, I. H. Thorseth, isotope fractionation by autotrophic bacteria with 3 different
and O. Tumyr, Microbial fractionation of carbon isotopes CO2 fixation pathways, Z. Naturforsch., 44c, 397 – 402,
in altered basaltic glass from the Atlantic Ocean, Lau Ba- 1989.
sin and Costa Rica Rift, Chem. Geol., 173, 313 – 330, Smith, D. C., A. J. Spivack, M. R. Fisk, S. A. Haveman, and
2001a. H. Staudigel, Tracer-based estimates of drilling-induced mi-
Furnes, H., H. Staudigel, I. H. Thorseth, T. Torsvik, K. Mueh- crobial contamination of deep sea crust, Geomicrobiol. J.,
lenbachs, and O. Tumyr, Bioalteration of basaltic glass in the 17, 207 – 219, 2000.
oceanic crust, Geochem. Geophys. Geosyst., 2, Paper number Smith, W. H. F., and D. T. Sandwell, Global seafloor topogra-
2000GC000150, 2001b. phy from satellite altimetry and ship depth soundings,
Furnes, H., K. Muehlenbachs, O. Tumyr, T. Torsvik, and Science, 277, 1956 – 1962, 1997.
C. Xenophontos, Biogenic alteration of volcanic glass from Staudigel, H., and S. R. Hart, Alteration of basaltic glass:
the Troodos ophiolite, Cyprus, J. Geol. Soc. London, 158, Mechanisms and significance for the oceanic crust-sea-
75 – 84, 2001c. water budget, Geochim. Cosmochim. Acta, 47, 337 – 350,
Furnes, H., I. H. Thorseth, T. Torsvik, K. Muehlenbachs, 1983.
H. Staudigel, and O. Tumyr, Identifying bio-interaction Staudigel, H., R. A. Chastian, A. Yayanos, and R. Bourcier,
with basaltic glass in oceanic crust and implications for Biologically mediated dissolution of glass, Chem. Geol.,
estimating the depth of the oceanic biosphere: A review, 126, 119 – 135, 1995.
in Volcano-Ice Interactions on Earth and Mars, edited by Stetter, K. O., G. Fiala, G. Huber, R. Huber, and A. Segerer,
J. L. Smellie and M. G. Chapman, Geol. Soc. Spec. Hypothermophilic microorganisms, FEMS Microbiol. Rev.,
Publ., 202, 407 – 421, 2002a. 75, 117 – 124, 1990.

21 of 22
 Geochemistry 3
Geophysics
Geosystems G banerjee and muehlenbachs: tuff life 10.1029/2002GC000470

Tarduno, J. A., W. V. Sliter, L. Kroenke, M. Leckie, H. Mayer, tion of oceanic crust, Chem. Geol., 126, 137 – 146,
J. J. Mahoney, R. Musgrave, M. Storey, and E. L. Winterer, 1995.
Rapid formation of Ontong Java Plateau by Aptian mantle Thorseth, I. H., T. Torsvik, V. Torsvik, F. L. Daae, and R. B.
plume volcanism, Science, 254, 399 – 403, 1991. Pedersen, Diversity of life in ocean floor basalt, Earth Pla-
Thorseth, I. H., H. Furnes, and H. Heldal, The importance of net. Sci. Lett., 194, 31 – 37, 2001.
microbiological activity in the alteration of natural basaltic Torsvik, T., H. Furnes, K. Muehlenbachs, I. H. Thorseth, and
glass, Geochim. Cosmochim. Acta, 56, 845 – 850, 1992. O. Tumyr, Evidence for microbial activity at the glass-altera-
Thorseth, I. H., T. Torsvik, H. Furnes, and K. Muehlen- tion interface in oceanic basalts, Earth Planet. Sci. Lett., 162,
bachs, Microbes play an important role in the altera- 165 – 176, 1998.

22 of 22