Biological and organic constituents of desert varnish: review and new
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Biological and organic constituents of desert varnish: review and new
hypotheses
Randall S. Perry *1 and Vera M. Kolb2
1
Department of Earth and Space Sciences, Astrobiology Center for Early Evolution,
Box 351310, University of Washington, Seattle, Washington 98195-1310
2
Department of Chemistry, University of Wisconsin-Parkside, Kenosha, Wisconsin 53141-2000
ABSTRACT
Desert varnish coatings are found on rock surfaces throughout arid regions of the world. Rock varnishes may exist on
Mars, as suggested by some observations on both Viking and Mars Pathfinder landing sites. There has long been a
debate as to whether varnish coatings are microbially mediated or deposited by inorganic processes. Dozens of bacteria
have been cultured from the surface of varnish coatings and recently the molecular ecology of varnish coatings have
been characterized using 16S rRNA techniques. Colonies of micro colonial fungus are associated with varnish coatings
but it is unclear whether bacteria or fungi are directly involved in varnish formation. Re-examination of a ten-year-old in
vitro experiment to grow varnish-using bacteria cultured from varnish surfaces provides some answers and is useful in
designing new lab experiments. Another alternative is the incorporation of microbial components into varnish coatings
either by complexation with metals or in association with clays or silica. For instance, polysaccharides found in bacterial
cell walls contain linear polymers of sugars that may be preserved in arid conditions when complexed with usual varnish
components such as calcium, aluminum, silicon, iron and manganese. Understanding the organic components of desert
varnish may help to resolve the question of the mechanism of formation of rock coatings, biomineralization processes,
and bacterial fossilization and how to detect past microbial activity on planets.
Keywords: Desert varnish, rock coatings, silicic acid, amino acids, DNA, Microcolonial fungi
1. INTRODUCTION
So much has been said about desert varnish that one wonders if there is anything else that can be said about the topic
without repetition of old ideas. Most of the approach to desert varnish so far has been reductionist, by necessity. Desert
varnish, however, is a complex phenomenon, an entire miniature Gaia, by itself, in which rocks, dust, bacteria and
microcolonial fungi (MCF), water, and desert sun all play roles, which have not been clearly understood.
There has long been a debate as to whether desert varnish is a result of inorganic or of biological processes. To date,
few explanations for formation have been suggested for inorganic mechanisms. Most of the ideas about formation have
been centered on microbial processes. Most of these have been put forth because microbes or their remnants have been
found in association with varnished rocks. Recently, it has been suggested 1 that the remnant organic constituents of
dead microbes may act in consort with inorganic processes. This process then, although involving organic compounds,
would be a non-biological process. However, since varnishes occur in diverse environments, it is entirely conceivable
that their compositions and mechanisms of formation are different. The central question remains: is there an underlying
principle for varnish or possibly all rock coatings?
*rsp@u.washington.edu; phone 1 206 543-6267; fax 1 206 685-2379First we review various biological hypotheses by others and ourselves, including our current findings of organic
constituents of coatings. We then present a novel mechanism for the formation process and then introduce a new
variable: the role of silicic acid in cementing varnish.
Alexander von Humbodlt, a German geographer, in 1799, described granite boulders among the cataracts near the
mouth of the Orinoco River in northeastern Venezuela as covered with a “smooth, black, and if as coated with
plumbago”. Strangely, we are still in an observational stage of the desert varnish theory. We know about the association
between MCF, bacteria and varnish, but we do not have a clear-cut cause-and-effect explanation. We suggest here a
possible connection between the mechanisms of biomineralization involving microbial exo-polymeric carbohydrate
substances and silicic acid, and a possible entombment of MCF, fungi, spores, and bacterial products in the rock coatings
via their polymerization with silicic acid. The hardening of varnish could be due to the additional crosslinking of silicic
acid or its sugar and other organic complexes, with ferric and manganese hydroxides.
1.1 BACKGROUND AND QUESTIONS
Desert varnish is widespread on Earth 2 and its existence on Mars has been proposed based on data from various
missions. 3 If it is present on Mars, is the process of formation similar to that on Earth? Do microbes cause the clay and
oxide rich, black-to-brown coatings of desert varnish? Are they precipitated inorganically, or is it something in between,
requiring a mix of inorganic processes acting on or with organic compounds? Even though we still do not understand
what the mechanism or mechanisms are, we do know that clays, oxides, detritus, and a cornucopia of humic substances
become incorporated into relatively hard varnish coatings (~6.5 hardness). The dominate chemical elements are Si, O,
and Al, with lesser and variable amounts of Fe, Mn, C, Ca, Na, K, N, P, Ti, Mg, S, Ba, and Cl. Microbes possibly
participate in this process either directly or indirectly and may play a role in manganese concentration. 4-6 Key questions
are: how is manganese concentrated into varnish coatings at levels far exceeding soils; what glues varnish components
together; what makes desert varnish almost as hard as quartz (~7 hardness); and what causes laminations and botryoidal
(Fig. 1) morphology?
10µm
Figure 1. Left image: desert varnish coated
rocks in the Mojave Desert near Baker, CA.
The dark shinny Mn rich varnish coating
covers rocks of different types making them
all appear black. Top image: SEM is of a
varnish coating from Death Valley, CA,
showing botyroidal growth form.
Typical manganese concentration in soils is 0.01% and in some coatings as much as 20 wt. % oxide. Concentrations
of this magnitude have led many researchers to look to microbial precipitation mechanisms. It is reasonable to test a
microbial mechanism, since the coatings exist in a biological world. 7 Even in arid regions, biofilms can form when
moisture is available. On Earth, no matter how arid the environment, water is still present to drive bio-chemical
processes, and where there is water, microbes find ways to flourish.
Many authors have described the physical characteristics of desert varnish. 8-12 Varnish is a coating and not a
weathering product of its substrate; therefore, the source materials derive primarily from dust deposits. Dust lands on
rock surfaces, some components are concentrated and become adhered or “glued” together. Unused materials need to be
removed in a continuing “conveyer belt” that allows for the removal, redepositon and concentration of mineralscomponents. This process eventually builds coatings up to 200µm thick or more as shown in Fig. 2. Sulfate and oxygen-
17 analyses 13 support this view and suggest that sulfate in the varnishes comes from the atmosphere and that biological
processes would have only a minimal effect on δ17O signatures.
Biologically mediated processes, however, have been proposed to explain the concentration and incorporation of Mn
and Fe in varnish relative to that in dust deposits. Several studies have found microorganisms or their by-products in
association with varnish. 1, 4, 6, 7, 11, 14-20 Some of these microorganisms are capable of oxidizing manganese. 4, 21
Organic chemical components have not yet been fully investigated. 1 However, studies of amino acids 1, 17 have
suggested that microbes and, possibly, Gram-positive bacteria are associated with varnish coatings, and DNA has been
extracted from the surface of rocks, soils and monuments. 22 A diverse
microbial ecology has been suggested for varnish by gene sequences of
varnish from Death Valley using 16S rRNA, amplified by polymerase
chain reaction (PCR) with primers specific for organisms in domains
Bacteria and Archaea. 18 There appears, however, to be no one single
bacteria or group of bacteria present in the DNA extracts.
2. GENERAL HYPOTHESES
Some researchers have invoked inorganic origins,9, 23 but most
suggested a biological involvement. 1, 6, 7, 15, 16, 20, 21, 24 It must be taken
into account that varnish forms in a natural environment and microbes
100µm and/or their products will have some influence on its formation.
Organics may also be added to varnish surfaces in aerosols or as
Figure 2. SEM of a thin natural varnish attached or complexed to dust particles.
coating on a rock substrate. While the mechanism of formation has remained a mystery, some
aspects, such as Fe and Mn accumulation, have been addressed. 14, 25
Adams et al. (1992) suggested a mechanism that involves bacterial and fungal chelators to concentrate Fe oxides and
oxyhydroxides. Fe2+ (uncomplexed) is unstable above pH ~5, which is typical of the soil, and thus precipitates. Fe is
concentrated 3-4 times in varnish over typical dusts. Based on experimental simulations, Jones (1991) suggested a
mechanism for the concentration of Mn in weak acids, such as carbonic (H2CO3). 25 Mn readily remains in solution as
Mn2+ and Mn3+ until pH reaches ~8.5 where it precipitates as insoluble MnO2. It should be noted, however, that
microbes might enhance Mn oxidation. 16, 19, 21 Many researchers have made the association of desert varnish with
bacteria and fungi, and causative relationships have been suggested but not proven. 1, 6, 7, 15-17, 19, 20, 24
3. HYPOTHESES AND MECHANISMS
3.1 The role of MCF and bacteria
The types of organisms that have been reported in the most extreme environments in which varnish occurs are
restricted primarily to bacteria and fungi. These organisms fall into the category of extremophiles in that they are
subjected to extreme summer heat and very low humidity. While bacteria have been cultured from surface swabs,4, 6, 10,
15, 16
they are rarely visible using scanning electron microscopy (SEM). 7, 18, 25 Bacteria may be less adapted than fungi in
hot arid conditions. 21, 26 Microcolonial fungi (MCF) have been frequently noted in association with varnishes in many
desert locations including Australia, the Sonoran, the Mojave, and Peru. 7, 10, 25, 27-29
MCF occurs as microcolonies, and therefore has been referred to as Microcolonial fungi. The colonies are easily
observed using an SEM (Fig. 3) and with a hand-lens on light colored rocks such as quartz, but are difficult to discern on
black varnish surfaces due to their black pigmentation. The dark color is primarily due to melanin. Some MCF are quite
heat resistant and can survive temperatures as high as 100°C for a period of two days.30 In situ respiration of varnish
chips with MCF was documented using 14C-labelled acetate. 7 MCF from some rocks have been found to incorporate
manganese within their colonies whereas no enhanced Mn was detected adjacent to the colonies. 28 We have observed
several examples of MCF degrading, becoming mineralized and apparently incorporated into varnish surfaces as did
Taylor-George et al. (1983). Both our lab and that of Anna Gorbushina at Oldenburg University have started molecular
investigations to identify different types of MCF.
Hungate et al. (1987) found 79 bacteria of which 74 were capable of oxidizing manganese from isolates obtained by
plating from the Negev Desert. The predominant genera were Bacillus, Geodermatophilus, Arthrobacter, andMicrococcus. A single manganese-oxidizing Actinomycete was also isolated. Similar genera have been cultured from the
Sonoran and Mojave Deserts but many of the strains appear to be different. Two groups that dominate the isolates from
the American Southwest are Arthobacter and Micrococcus. Also a single Bacillus sp., Planococcus sp, and Hyphomonas
sp were identified as manganese oxidizing bacteria. All of these microorganisms were Gram-Positive. Gram-negative
microorganisms were reported by Dorn and Oberlander (1981) when they suggested microorganisms that were similar to
Metallogenium and Pedomicrobium were responsible for varnish coatings.
3.2 Spore mechanism
Spore forming bacteria on exposed rock surfaces exhaust nutrients as
water becomes scarce in summer and sporolate. Cellular products from
the mother cells eventually lyse and contribute organic polymers to the
rock surface. The remaining spores concentrate and become encrusted in
minerals, particularly those of Ca2+ and Mn4+. The mineral encrustations
provide protection from harsh conditions and impart heat resistance.
Spores are generally heat resistant to ~100°C. However as temperatures
exceed 50°C, some spores such as B. Subtillus and B. Megaterium begin
200µm
to demineralize. The demineralized products rich in Ca2+, Mn4+, other
oxides and organic compounds might then be deposited on the rock
surfaces. Surviving spores might deposit byproducts to rock surfaces
during germination. The spore demineralization products might then
combine with extracellular polysaccharides, detrital clays, or oxides to
form varnish coats. A unique component of spore coats is dipicolinic
acid. Finding dipicolinic acid in varnish coats would support this
hypothesis.
3.3 Interaction of bacteria and saccharides with the natural surfaces
100µm The most likely part of bacteria to interact with rocks is the outer part
of the cell wall that contains oligosaccharides from peptidoglycans, but
Figure 3. Microcolonial fungi on varnished rock
surfaces from Bishop, CA.
also other saccharides, such as teichoic acids and related sugar-lipids.
The bacteria, upon attachment to the surface also produce slimy adhesive
substances that are predominantly exopolysaccharides (EPS). 31 It is likely that bacterial polysaccharides will interact
with the natural surfaces in a process that is probably facilitated by their initial complexation with metals that are found
on these surfaces. The complexation may be followed in some cases by a redox-type reaction. It is known that
peptidoglycans are highly interactive with dissolved metal ions. 31 This may be the beginning of a sequence of reactions
that lead to the cementing of bacteria to the surface. An additional pathway would be via complexation with silicates, of
a type described by Kinrade et al. (1999, 2001) for polyols and sugar acids.
Peptidoglycans are complex polysaccharides found in the bacterial cell walls. Peptidoglycans contain linear polymers
of two alternating sugars, N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM), that are cross-linked with the
short peptides. These peptides are composed of some common amino acids, as well as some unusual ones, such as D-
glutamic acid, diaminopimelic acid (DAP), and D-alanine. 32 The unusual amino acids have been suggested as possible
biosignature of the bacteria in desert varnish. 1 Another possible candidate for a biosignature within peptidoglycans
would be the peptide inter-bridge composed of five glycines, which is found in some gram-positive bacteria. The peptide
cross-links in the peptidoglycans may protect the sugars from decomposition and thus enable them to serve as
biomarkers.
One can reasonably ask if a simultaneous finding of these unusual amino acids, pentapeptide bridge, and the NAG
and NAM sugars or their transforms, could indicate the remnants of the cell walls of bacteria. This would be of extreme
importance in the identification of bacterial fossils, and for the understanding of the processes by which various biofilms,
coatings, and rock varnishes form on the natural surfaces and undergo mineralization over time.
In general, sugars have not been studied as a biosignature. There is ample reason for this. Common sugars, such as
glucose, ribose, arabinose, or fructose, contain aldehyde or keto groups in conjunction with the hydroxyl groups that
make them very chemically sensitive. Such sugars are rapidly destroyed under the basic conditions. They isomerize
under both acidic and basic conditions. Isomerization causes racemization of the optically active centers, and an
eventual destruction of the molecules, 33 thus preventing their use as biomarkers.However, some sugar derivatives that are devoid of the aldehyde and keto groups, notably sugar-related acids and
alcohols, are more stable and thus have been isolated from the Murchison meteorite. 34
Sugars may be more stable under the arid and semi-arid conditions under which most common rock varnishes are
formed. Sugars in general, such as carbohydrates and polyols, sugar acids, carbohydrates with nitrogen or other
electron-donor atoms, oligomeric- and macromolecular carbohydrates, make a variety of stable complexes with metals,
such as calcium, aluminum, iron, manganese and others, 35 that are commonly found on natural surfaces. Currently we
are investigating the possibility of sugars complexing with components of desert varnish in laboratory simulations.
3.4 Photochemical processes
One aspect of the natural environment that has not been systematically studied in the laboratory setting is how light
might influence chemical processes of varnish formation. Reductive dissolution of metal oxides from attached soil
particles would probably alternate with re-oxidation and precipitation of the dissolved metals in a cyclic fashion driven
by the daily wet-dry regime of the desert climate. At dawn, during much of the year, a thin film of water forms on rock
due to the condensation of dew, wetting aeolian soil particles. Photochemistry begins as soon as the sunlight hits the
wetted rock surface, reducing iron and manganese oxides by oxidizing organic compounds, such as α-carboxylic acids.
Oxalic acids are readily available from fungi and lichens. On the inorganic side, these reactions would lead to the
solubilization of metal elements, yielding ferrous iron and manganese (II) in example. The reduced metals could stay in
solution via organic chelating agents. At the same time, the presence of organic acids including amino acids also causes
dissolution of silicates, providing a source of dissolved Si. On the organic side, the photochemistry oxidizes the α-
carboxylic acids to CO2, reducing the amount of organic content on the rock surface. In the case of glutamic acid, this
photo degradation leaves a diagnostic byproduct, γ-amino butyric acid (GABA), which has been found in desert varnish.
As the dew evaporates and the rock dries out, the dissolved species begin to oxidize and precipitate, a process likely to
result in new mineral phases.
3.5 The Role of silicic acid
The role of silicic acid in desert varnish formation has not been explored previously. We suggest that silicic acid is of
primary importance not only for the formation of desert varnish, but also for some other rock coatings and glazes.
Silicic acid, Si(OH)4 , is the principal product of dissolution of silica, SiO2, which is quartz or amorphous silica. The
solubility of SiO2 is very small at pH’s that are less than ~ 9. At the high pH’s, concentration of silicate ion is increased.
The solubility of silicic acid is enhanced by the presence of amino acids and other organic molecules including
carbohydrates. 36 Silicic acid is prone to polymerization via condensation of the silanol groups, SiOH, a process in which
water is eliminated (Fig. 4), and silica gel is formed (Fig. 5).
pH > 9
nSiO2 + nH2O nSi(OH)4 n(HO)3Si-O- + nH+
Silica Silicic acid pH < 9 Silicate Ion
OH OH OH OH
H2 O
HO Si OH + HO Si OH HO Si O Si OH
OH OH OH OH
Figure 4. Formation of silicic acid, silicate ions, and disilicic acid.
The production of silicic acid under the desert varnish conditions is dependent on the presence of water, organic
materials, and pH. Water can remain on surfaces for many days in cool overcast weather or may be evaporated rapidly
on hot sunny days or in the low humidity conditions present in arid regions. If a significant amount of water relative to
dusts is present, then the variable but usually small amount of aeolian deposit does not significantly alter the pH over
hours to days. If new water is not added then evaporation begins and the pH level is raised as salts and clays buffer the
solution. Many chemical processes occur in water-filled depressions including those of complexation and photochemical
as mentioned previously. Small amounts of silicic and disilicic acid, ((HO)3Si-O-Si(OH)3), are formed. When water
evaporates, pH rises, and the solubility of silica increases and more (di)silicic acid is formed.37 When water evaporates,silica gels are formed by condensation. Eventually in deserts, the gels, especially at the surface and outer layers, are
dried and dehydrated in low humidity and are exposed to a hot sun where surface rock temperatures often rise to 70ºC or
more. During baking, progressively more hydroxyls are driven from the gels resulting in hard coatings.
Silicic acid can form a variety of
H
OH
OH complexes with ions and organic
OH
O molecules. Examples are
OH
Si O - mucopolysaccharides, glycoproteins that
O O O are enriched in hydroxyl amino acids
Si Si O
-
O O
Si O OH
(serine and threonine), glycine, aspartic
O
O
O Si
and glutamic acid 38. In addition, silicic
Si
Inside O acid is expected to form organic silicate
O Si OH
O Si complexes (Si-O-C) with oxyanion centers
Si
O derived from cis-1,2-diols that are fixed at
HO Si
O
O
OH ca. 0.26nm. Candidates include some
oxyanions of sugars, unsaturated
O OH
Si O Si polyhydroxy compounds, catechols (1,2-
HO Si
O O diphenols), and other compounds with
O Si
OH rigid structures and a correct “bite” that
H
OH
OH matches the O-Si-O angle and thereby
HO makes a stable complex. 39 Flexible sugar
OH
related substances, such a polyols and
sugar acids, also make Si-O-C complexes with
Figure 5.Amorphous hydrated silica. (after Mann, S., silicic acid if they possess at least four hydroxyl
Biomineralization p. 15. groups in a particular stereochemical arrangement.
40-42
From these examples it is easy to see how
various organic compounds that are present in the soil, or blown to the varnish surface, or come from the bacteria and
fungi that are present, or their remains, can make Si-O-C complexes with silicic acid and contribute to the crosslinking
and hardening of the silicate polymers that are formed. Significantly, silicic acid also makes Si-O-metal complexes, such
as the Si-O-Fe complex with ferrihydrate. 43 Detrital clays are incorporated into varnish coatings. 9 However, typical
clay components of Al, Mg, Na, and associated organic compounds also can complex readily with silicic acid. 44This
means that not only organic substances, but metals as well, can participate in polymerizing, crosslinking and hardening
(by elimination of water) of the silicic acid.
This proposed mechanism for varnish formation can be applied generally to most subaerial coatings, including silica
glazes 45-47 and various iron (red) and manganese (rich dark) coatings from arid and semi-arid worldwide locations.
Silica glazes from Peru and Hawaii 45 resemble desert varnish in their luster and hardness. They are thin and often
exhibit growth patterns similar to varnishes, but otherwise do not resemble dark varnish coatings that have additional
mineral, detrital and organic components from their local environment.
4 METHODS
4.1 Molecular ecology sample collection
A rock sample from Death Valley was brushed with a sterile brush to remove loosely attached soil in situ. Varnish
was removed from the rock surface using a Dremel tool and collected using a sterilized brush and brushing the contents
onto sterile paper. The fine powder was placed in glass tubes and transported to the lab and placed into a -60ºC freezer.
The powder remained unfrozen for 7 days. The rock was brought to the lab and more varnish was removed and the
DNA was extracted within one day. Soil from the area near the rock was also collected.
4.2 DNA extraction, PCR, cloning, and sequencing
DNA was extracted from 500 mg of varnish or surrounding soil using a Fast Soil DNA extraction kit (Bio101) and
Beadbeater (Savant). Approximately 2 µg and 6 µg of DNA were obtained from 500 mg of varnish and soil,
respectively. Polymerase Chain Reaction (PCR) to amplify16S ribosomal RNA (rRNA) genes was performed using
DNA from soil or varnish as template with primers 27f and 1492r used in standard 30-cycle PCR with Taq polymerase
and an annealing temperature of 50ºC. For amplification of Archaea 16S rRNA genes, primers 20f and 1492r were used
and PCR amplification was performed using the same technique as for the Bacteria.4.3 Library construction
Screening and sequencing Clone libraries of bacterial- and archaeal-specific PCR product from desert varnish were
constructed by cloning product using a TOPO TA cloning kit (Stratagene) and transforming into E. coli. Resulting
clones were screed by amplified ribosomal DNA restriction analysis (ARDRA) using restriction enzymes RsaI and MspI.
Clones with similar restriction patterns in both digests were designated to a common operational taxonomic unit (OUT).
Full length (single strand) sequence was obtained by sequencing the same primers used for PCR amplification. Only
subsets of the screened clones have been sequenced thus far.
4.4 Phylogenetic Analysis
Full length sequences (DV Bact-X and DV Arch-X, with X representing the clone number) along with high BLAST
hits were initially aligned and checked for chimeras using the Sequence Alignment tool and Chimera Check tool,
respectively, at the RDP website. Sequences were then manually aligned using the program GeneDoc, not using any
positions that could not be unambiguously aligned. Phylogenies were inferred with the program TreeCon, 48 using the
neighbor-joining method with a Kimura 2-parameter base substitution model to develop trees. One hundred bootstrap
analyses were done for each tree. Out groups used were Aquifex pyrophilus for the bacterial 16S tree and Methanococcus
jannaschii for the Crenarchaeon 16S tree.
4.5 Scanning electron microscopy (SEM) and energy dispersive microanalysis system (EDAX)
SEM analyses were done at Pacific Northwest National Laboratories (PNNL), a Department of Energy facility in
Richland Washington. SEM imaging was performed using a LEO 982 Field Emission Scanning Electron Microscope,
which is an ultra-high performance scanning electron microscope with a resolution 1 nm at 30 kV and 4 nm at 1.0 kV.
The high resolution is achieved using a Schottky field-emission source, a beam booster that maintains high beam energy
throughout the microscope column, an electromagnetic multihole beam aperture changer, and a magnetic field lens. The
beam path has been designed to prevent crossover of beam electrons. These features result in reduced chromatic
aberration, improved beam brightness, and little beam energy spread. The Leo 982 has two secondary electron detectors:
below lens and in-lens for high resolution imaging. The backscattered electron detector is solid state and is optimized for
short working distances.
EDAX was performed on an Oxford ISIS used for chemical analyses. A SiLi detector, having 128 eV resolution, is
capable of light element analyses, elemental mapping digital imaging, microscope automation, and can combine
compositional information with a secondary electron image in a software package called CAMEO. Samples were coated
with gold, platinum, or carbon depending on what elements were being analyzed.
4.6 Transmission electron microscopy (TEM) with EDAX
TEM-EDAX: High-resolution TEM analysis was carried out on a Jeol JEM 2010F microscope with a specified point-
to-point resolution of 0.194 nm. EDAX is an Oxford Link system with ISIS analytical software.
4.7 Time-of-flight secondary ion mass spectroscopy (TOF-SIMS)
TOF-SIMS measurements were carried out on a Physical Electronics TRIFT II TOF-SIMS using a 69Ga+ source in a
high spatial resolution mode. In this mode, a 25 kV, 60 pA unbunched primary ion beam with a 30 ns pulse width is
used. Although the mass resolution in this mode is approximately m/∆m = 1000, the primary beam diameter ofThe instrument has a 16-element multichannel detection system. The X-ray beam used was a 70W, 100um diameter
beam that is rastered over a 1.4 mm by 0.2 mm rectangle on the sample. The x-ray beam is incident normal to the
sample and the x-ray detector is at 45° away from the normal. The survey scans were collected using a pass energy of
117.4 eV. For the Ag 3d 5/2 these conditions produce FWHM of better than 1.6 eV. The high-energy resolution data
was collected using a pass energy of 23.5 eV. For the Ag 3d 5/2 these conditions produce FWHM of better than 0.75
eV. The collected data were referenced to an energy scale with binding energies for Cu 2p 3/2 at 932.62± 0.05 eV and
Au 4f at 83.96.0± 0.05 eV. A BaO cathode was used as a source of electrons for neutralization providing 1 eV energy
electrons at 21 uA. Low energy Ar+ ions were also used for specimen neutralization.
4.9 Stable Isotopes: δ13C and δ56Fe
Isotope analyses were performed by elemental analyzer-continuous flow isotope ratio mass spectrometry
(EA/CFIRMS), using a Carlo Erba NC2500 EA interfaced through a Finnigan CONFLO II to a Finnigan Delta XL mass
spectrometer. Sample isotope ratios were normalized in each run to the values obtained for an organic standard with
known isotope ratios calibrated via sealed-tube combustions versus NBS-19 at δ13 C = 1.95‰ vs Vienna PeeDee
Belemnite(VPDB) and for δ15 N = 0.0‰ vs air nitrogen. Precision in this system averages ± 0.12‰ for organic standards
and homogenous natural samples. Accuracy, as measured by including repeats of a natural sample of known isotopic
ratio in each run, was ± 0.10‰. All isotope ratios are expressed in delta notation, or parts per thousand deviation from
the VPDB standard, where: δ13C ={[(13C/12C)sample/(13C/12C)VPDB ] –1}× 1000
Iron isotope compositions were measured using either a Micromass Sector 54 thermal ionization mass spectrometer or
a Micromass IsoProbe, 49 a multiple-collector inductively coupled plasma mass spectrometer with a magnetic sector.
Isotope analyses by thermal ionization mass spectrometry (TIMS) used a 54Fe-58Fe double spike to correct for
instrumental mass bias. External precision (1 SD) of TIMS isotope analyses is ±0.1%/ amu (i.e., 56Fe/54Fe is ±0.2%), as
determined by replicate analysis of samples and ultrapure Fe standards.
The IsoProbe is a single-focusing multiple-collector inductively coupled plasma mass spectrometer that uses a
hexapole collision cell to thermalize the ion beam so that the ion energy spread is reduced to ~1 eV. Additionally, the
hexapole collision cell, with a mixture of Ar and H2 gas, eliminates or minimizes argide ions that are isobaric with Fe.
Fe-isotopic ratios are in conventional per mil notation: δ56Fe ={[(56Fe/54Fe)sample/(56Fe/54Fe)E-M ] –1}× 1000
5. RESULTS
5.1 Microbial and molecular ecology
We present here the first results know to us of culture-independent techniques for assessing microbial diversity in
desert varnish coatings from the Mojave Desert. This study indicates that there are a wide variety of prokaryotic
microorganisms on or in varnish surfaces. Scanning electron microscopy (SEM) was used to determine if the presence of
microbes on the surface of the varnish coatings. Few bacteria were observed, suggesting that the DNA may be derived
from inside the coating, but not proving that the product was not a result of enhancing bacteria from the coating. The
sequencing suggests a diversity of microbes, while the previous culture dependent studies and amino acid studies suggest
primarily Gram-positive bacteria. Bacteria included Alpha and Beta Proteobacteria, one Gram-Positive, Acidobacteria,
Deinococcus/Thermus, Cyanobacteria Chloroplasts, and Green non-sulfur bacteria. Six clones of non-thermophilic
Archaea were identified but no representatives of the thermophilic Crenarchaeota were found. There are no reports
known to us of cultured representatives of the non-thermophilic Crenarchaeota.
Microcolonial fungi have been found on varnish coatings in all samples that we observed. The colonies are of variable
size, typically ~ 100µm and black in color.
5.2 Search for dipicolinic acid
Dipicolinc acid is a chemical signature for spores. Utilizing TOF-SIMS, we failed to find this acid in samples from
the Mojave Desert. This suggests that sporolating bacterium or their spores were absent. However, the presence of
bacteria could differ significantly with climatic conditions and location. The samples tested were collected during winter
from Grimes Point, Nevada and Death Valley, California. We would expect that the winter months would be a less
hostile microbial environment and thus would be conducive to bacterial growth. In laboratory experiments (vide infra)
many spores were seen with the SEM (Fig. 6), however, spores have been few and far between on natural varnish
samples we have viewed.5.3 Laboratory invitro studies
Understanding desert varnish genesis ultimately should involve testing a formation theory and creating varnish
coatings in the laboratory. Previously, two separate laboratories have examined the in vitro production of varnish using
pure cultures of microorganisms. Krumbein and Jens (1981) isolated, from rocks with desert varnish, several strains of
bacteria, fungi, and cyanobacteria that were able to precipitate iron or manganese. They prepared agar plate media for the
growth of microorganisms and included in the media (a) rock chips from the interior of rocks, (b) rock particles that had
been acid cleaned, or (c) acid cleaned quartz sand grains. These particles protruded from the surface of the agar. The
plates were then inoculated with the isolates of bacteria, cyanobacteria, or fungi. The microorganisms grew both on the
agar surface and on the particles, and deposited iron and manganese oxides in the vicinity of growth. When the agar
dried, a rock varnish-like film was produced. Rock particles were covered by iron and manganese oxide deposits,
especially in the vicinity of fungal hyphae and cyanobacterial cells. The mineralogy of the films was not determined,
leaving open the question of whether or not
Table 1. Bacterial strains analyzed in vitro study
varnish was formed. Dorn and Oberlander
(1981) performed lab experiments to make Bacterial Strain Source Assigned Genus or Gram Mn Oxidizer
varnish and identified “Metallogenium”-like ________________________________Description_______Reaction_______________
or “Pedomicrobium”-like bacteria, in a _
liquid medium, and reported that deposits N-15 Negev Geodermatophilus + +
were precipitated. 15 85-9B Australia Bacillus + +
Palmer, Adams and Staley initiated a 85-17 " Bacillus + +
laboratory study in 1992 (unpublished) with 85-24B Mojave small rod + N.D.
the objective to grow varnish resembling
natural varnish. Frosted glass squares 1 cm2 were sterilized and placed in mounts in sterile petri dishes. Slides were then
inoculated with bacterial strains listed in Table 1 in various combinations with soil extract medium (Table 2). Soil
extract (SE) was prepared using soil from an area of desert-varnished rocks near Phoenix, Arizona. 7 Five hundred grams
of soil was mixed with 500 ml of water, autoclaved at 121ºC for 1 hr, and filtered through Whatman #2 filter paper. The
medium was composed of 0.005% sodium acetate,
0.02% MnSO4·H2O, 0.02% FeSO4·7H2O, 0.0001% Table 2. Soil Extract (SE) media and inocula used in Experiments
cupric citrate, 6.5% bentonite in soil extract, final
pH 6.8 after autoclave sterilization. Slides were Medium Inocula
In vitro Experiment
allowed to dry, washed gently with 0.01 N
1 SE + Mn None
NaHCO3 and placed in a desiccator to dry. After 2 SE + Fe None
drying, 0.05 ml soil extract was added to each. 3 SE + Mn + Fe None
Usually about 24h was required for the nutrients to 4 SE + Mn N-15
dry, after which the slides were washed and dried 5 SE + Fe 85-9B
again as described above. Uninoculated sterile 7 SE + Fe + Mn 85-17, N-15, 85-9B, 85-24B
controls received the same treatment. Iron, 12 SE + Mn None
manganese, clays, other varnish components, and
nutrients for the activity of bacteria were available in the soil extract. To the soil-extract medium either FeSO4·7H2O,
0.005% or MnSO4·H2O, 0.005% or both together was added. The pH was adjusted to 6.9. Sterile montmorillonite
(0.05ml of 1% suspension) was added at intervals to supplement the clay in the soil extract. The procedures described
above (innoculation, drying, washing, and drying) were followed over 26 months with two 6-month periods when the
slides received no treatment. The slides were heated to 80ºC for 10-12 h and reinoculated 7 times over the 26-month
period. There were 72 cycles of feeding, drying, washing, and drying. The use of sodium bicarbonate between
applications of media/bacteria possibly raised the pH (the ending pH increased to ~9.4). It is possible that the pH
increase precipitated the coatings, as all samples analyzed showed Ca in the deposits. Fig. 5 shows large amounts of Ca
in Inv 7 (inoculi had 4 strains of bacteria) and a spheroid deposit (Fig. 4, Inv 12) enriched in Ca and somewhat enriched
in Mn and containing Mg, Na and lesser amounts of Si, P, S, and Cl. Inv 12 had Mn and SE but no bacteria added. Light
and humidity were not controlled during the ~ two year experiment.
We did not feel that it would be productive or practical just to shuffle the variables and/or attempt to redo the
experiment. Consequently we evaluated the coatings formed in one of their experiments using SEM with EDAX, and
analyzed 13C stable isotopes (Table 3) of the precipitates to attempt to determine if the coatings were of biological or
abiological origin. We report here for the first time our results. The choice of which experiments to evaluate was not
arbitrary but dictated by samples that remained available for re-examination and provided to us by Palmer et al.SEM images made by us in 2002, ten years after the original experiment, are shown in (Fig. 6). For comparison see
Fig. 1 for a SEM image of a natural varnish surface. EDAX analyses of their in vitro coatings are shown in Fig. 5, along
with a spectrum from a natural varnish surface from Bishop CA.
All deposits examined from the
original experiment have Fe3+
precipitates. On some of the slides Fe
Inv1 Inv4 oxides are probably derived from soil
extracts and/or other additions rather
then bacteria precipitates, as they were
present in controls that contained no
bacteria (Fig. 7, Inv. 12). Mn was
variable. Inv #2 (Fig. 4) shows no
10µm 10µm detectable Mn where SE and Fe were
used but no bacteria were added. EDAX
of Inv 5 with the same Fe and SE
additives, but with a Bacillus added
Inv 5 Inv showed a Mn precipitate. Inv 4, 5, and
7 show bacterial spores while none
5 were observed where no bacteria were
added (Fig. 4, Inv 1 and 12). As noted
above, several of the spectra show
substantial Ca, while S, Ti, Cl, K appear
10µm 2µm in varying amounts. Si, Al, Mg, and Na
appear consistently.
5.4 Fe isotopes
Inv Inv 12 Results of δ56Fe value for a varnish
scrapping from Death Valley, CA was
7 0.56 while nearby soil was 0.75.
Although these results are preliminary
and much uncertainty still surrounds the
10µm 20µm use of 56Fe as a biological indicator, the
finding of δ56Fe values in the positive
range suggests that Fe is possibly
Figure 6. SEM images of coatings formed in the experiments. Inv1 and 12 have no
precipitated via inorganic mechanisms.
bacteria added. Inv 4, 5, and 7 all have bacteria added (see Table 2), and bacterial
spores are present. The spheroids are Ca rich in Inv 12 ( see Figure 7).
6. DISCUSSION
6.1 Significance of the amino acid distribution
A recent study of amino acid hydrolyzates from varnish coatings in the Mojave and the Sonoran deserts isolated 13
amino acids that suggested bacteria were present or preserved in coatings. Protein amino acids cysteine and tryptophan
were not detected, while two non-protein amino acids, β-alanine and γ-amino butyric acid, were found. Two amino
acids that are components of peptidoglycan were also found, D-alanine and D-glutamic. The finding of L-lysine, D-
alanine, D-glutamic but not diaminopimelic acid is characteristic of Gram-positive bacteria supporting previous culture
dependent studies. 4, 6 It seems plausible that Gram-positive bacteria would be more likely to be preserved 50, 51 than
Gram-negative bacteria and the amino acid study of Perry et al. (2003) and that of Nagy et al. (1991) supported this
view. The study suggested that the “relatively high abundance of labile amino acids serine and threonine, along with the
lack of D-alloisoleucine, implies that the amino acids in desert varnish are possibly less than 200 y old assuming the
amino acids are homogeneous throughout the varnish rather than on or near the surface.” However, serine and threonine,
along with glutamic and aspartic may also form complexes with amorphous silica 52. It is the ability of these amino acidsto form complexes with the hydroxyl groups of silicic acid that provides an alternative explanation for their presence.
This is true particularly for serine and theronine, since they are not generally stable in the natural environment.
6.2 Significance of the silicic acid hypothesis
The process of formation of desert varnish, silica glazes and perhaps many other rock coatings could be silicification
via dissolution of aeolian silicates (and substrates), formation of small quantities of silicic acid (~ 5-130 ppm near neutral
solution and increasing with higher pH 37), formation of primary particles of condensed silica and their subsequent fusion
by gelling. The shrinking and drying process might account for botryoidal and lamellar growth forms in varnish. 53 If
bacteria or fungi are present, they may contribute to the weathering process of minerals and may produce more silicates.
Bacteria and fungi may eventually become silicified and incorporated into the coatings.51, 54 One would expect a diverse
genetic imprint in coatings. The molecular analyses of microbial communities present in varnish suggest great
diversity.18 The incorporation and complexation of amino acids with silicic acid favors serine, threonine, glutamic and
aspartic acid, 38, 52 but may include other amino acids. 36 These four amino acids were found in significant quantities in
varnishes from the Southwestern U.S.1
The surface zone where minerals, dusts, aerosols, water and microbes interact is complex. Thus the contributions of
the other processes must be taken into account, such as chelation processes of bacteria (via catechols) and fungi (via
hydroxamates), photochemical processes, oxidation and reduction of oxides by microbes, and other inorganic
mechanisms. Although the natural environment where varnish forms is dynamic and complex, we conclude that silicic
acid is possibly an important cementing agent which may incorporate other elements frequently found in varnish
coatings by complexation or entombing. These include but are not limited to Al, Fe, Mn, C, Ca, K, Na, N, P, S, Mg, Cl,
and Ba, as well as organic molecules, detritus, clays, and microbes when they are present. DNA might also be preserved
in silicic acid or possibly complexed with oxides entombed. Limiting factors in coating formations are amounts of water,
since too much water probably washes solutions from the surface and removes silicic acid
that is the product of the slow process of dissolution of silica. The small amounts of silicic Table 3 XPS atom
acid brought into solution supports the observations that varnish forms slowly. 12, 55 The %
rates of mineral leaching and oxide concentration and silicic acid deposition may be related after 5nm
to periods of wetness suggesting why varnish might be used as a paleoclimatic indicator as sputter
shown by Liu (2002). C 31.33 10.76
N 2.15 1.36
Continuing efforts to prove the validity of this hypothesis are necessary but intial O 47.16 56.34
experiments using XPS (Table 3) and TOF-SIMS show that Si is enriched in outer layer of Na 0.52 0.73
a black varnish from the Mojave Desert. This suggests the possibility of a climate Mg 1.05 2.19
controlled selective enrichment of elements in the silica. While substantial C was found Al 4.18 6.39
after sputtering the surface to remove 5nm of material, it was reduced from 31.33 atom% to Si 10.21 14.27
10.76 atom%. Nitrogen however was proportionally reduced less from 2.15 atom % to 1.36 K 0.89 0 .93
atom %. The presence of C and N might be due to organics including amino acids. Other Mn 1.01 3.34
elements were proportionally increased after sputtering.
6.3 On making varnish in vitro
The success of making varnish in vitro critically depends on the experimental protocol. While most variables involved
in varnish formation are known, the theory of varnish formation that would link these variables properly is still
underdeveloped. Einstein said, “Only theory can tell us which experiments are to be meaningful”. Unless we understand
the theory, the only other approach for the experimental design left would be a reductionist approach in which we would
test the importance of each individual variable separately. Due to the large number of variables that we know of, such an
approach is not practical. An intermediate approach in the future, would be to pick out some variables that are crucial,
neglect the others, and investigate them in an experimental setting that would be an approximation of nature.
The experiment of Palmer et al. was originally undertaken to see if coatings similar to desert varnish could be
produced in the laboratory using cultures of bacteria isolated from desert varnish. The coatings produced did not,
however, texturally resemble natural varnish coatings and were not as hard. The original deposits were soft, and weakly
adhering. Metals were concentrated in varying amounts as shown in (Fig. 7). In their experiment, microbes may be
implicated in many aspects of these processes by concentrating iron and manganese from soil additives or by
transforming clay precursors to clays and by producing polymers which may enhance packing of tiny particulates 56, 57
thereby aiding adhesion to rock surfaces 58. EDAX spectra did show that Fe was concentrated from the soil (Fig. 7-Inv1),
since no Fe was added other than that present in the soil. In Fig. 7-Inv 4 Fe appears to be slightly more enhanced even
though no Fe was added but bacteria was added. The original experiments did indeed produce coatings, but they weresoft and did not prove the involvement of bacteria. While coatings were produced, simulating “natural” conditions more
closely might have resulted in better adhesion of coatings.
Similar soft-coatings to those produced in the
experiments have been reported from nutrient-
limited caves 59. Sulfur, iron, and manganese-
Inv1 Inv5 oxidizing bacteria were found to generate
considerable acidity, dissolving cave walls.
Microbial induced mineralization included
silicate, clays, iron, and manganese oxides.
Inv2 Inv7 In the desert, dust settles on surfaces; some
material is retained and concentrated, and other
material is removed and thus not incorporated
into varnish coatings. It is interesting to note the
Inv3 Inv12 presence of iron (Fig 7 Inv 1,4). The iron must
Spheroids have been derived from the SE in the absence of
microbes.
Future lab studies should take into account
Natural Varnish both heat, light and the removal of
Inv4 From Bishop California unincorporated silicates. Varnish forms in the
presence of heat and light and its formation
requires water and is thus a solution deposit.
Figure 7. EDS spectrum of invitro coatings and a natural varnish coating from Bishop, CA. The source could be dew, rain, or snow.
Presumably dew would be available in varying
degrees nightly, while rains and/or extended wet
periods would be intermittent. Abiological mineral reactions may be essential in producing varnishes. For instance, the
effects of light and heat were not part of this experiment and certainly varnishes in nature are exposed to photochemical
process and temperature variations. Adding to the problem are the consortia of species possibly involved, each species
having a small role or possibly a symbiosis between two microbes, such as a fungus and a bacteria. Alternatively, it may
be entirely unimportant as to which microbes are present, as the extracellular products and decaying cells react with
metals, clays, water, light and heat to form coatings.
The results of the 13C stable isotopes show that the greatest fractionation was obtained for the Inv 1 was δ13C –25.45,
while Inv 5 was δ13C –22.92 where no bacteria were present. This also suggests that the in vitro coatings produced by
Palmer et al. were probably not of biological origin.
While the process may be more complex, some studies have emphasized microbial properties, which could be
important relative to this problem. The sorption of metals by bacterial cells 14, 60 and the production of high affinity iron
chelators and siderophores produced by bacteria and fungi, may be of special relevance 14, 61. These activities suggest
additional mechanisms for mobilizing iron from dust, and fixing iron and manganese to surfaces. In unpublished
research, we have found that when FeIII, MnIV, montmorillonite and siderophores (2,3-dihydroxybenzoic acid and
desferoxamine mesylate) are added to surfaces, with and without montmorillonite, coatings resembling varnish in color
and texture are produced. Of particular interest is the fact that these coatings adhere to surfaces more tenaciously than the
fragile coatings reported in the in vitro experiments of Palmer et al. We do not consider these siderophores-coatings
rock varnish in the true sense because they are probably higher in organic content and different mineralogically.
However, siderophores or other organic substances may contribute to the texture and binding of the coating to the
substrate.
There are unanswered questions as to whether the precipitates formed were made by periodically washing with 0.01
N NaHCO3. During drying, the residual bicarbonate solution will lose CO2 and become progressively more alkaline, and
consequently, could cause metal oxide deposits. The pH of the solutions applied during the experiment started at near
normal but resultant surface pH did become progressively more alkaline. Dorn (1989) records varnish having a pH near
9.0 in a series of varnish samples, while soil in the vicinity had a pH of 9.562. The highest pH, in that experiment, of 9.4
was recorded at the end of the experiment, but is not out of line with natural conditions. Oxide precipitation for iron is
favored above pH 5.5, but biological iron oxidation can occur at circumneutral pH. 63 Manganese can remain in solution
and begins to oxidize at a pH above ~8.0. Microbes may, however, facilitate the oxidation process at lower pHs.
Our analyses of 13C/12C isotopes might have answered the question as to whether the washing by bicarbonate caused
the precipitates. If the controls without bacteria would have shown a non-biological fractionation, then it might bepossible to conclude that the precipitates were biologically mediated. It does seem possible then that the bicarbonate
solution was at a higher pH and eventually raised the the pH of the coatings so that an inorganic autocatalitic oxidation
took place, rather than a microbially induced precipitation.
We found during the re-examination of Palmer et al. experiment and the resultant coatings to be valuable in designing
new laboratory experiments. Currently, we have on-going laboratory experiments utilizing involving silicic acid,
carbohydrates, amino acids and microbes.
6.4 Microbial and molecular ecology
DNA was extracted from the surface of rocks 22 and from varnish coatings 18. The 16S rRNA gene sequences were
amplified by polymerase chain reaction (PCR) using primers specific for organisms in domains Bacteria, Archaea and
Eukarya. A large variety of bacteria were found and representative sequences were used to construct a phylogenetic tree.
Six Archaea were also identified and sequenced. Bacteria used in this experiment are currently being sequenced and will
be compared to DNA found in coatings. The large variety of organisms found suggests that many different bacteria or
extracellular products are available, and may contribute to varnish formation or become incorporated in varnish coatings.
Nagy et al. (1991) described more fully than previously had been done, the physical and structural function that
bacteria and fungi may contribute to desert varnish. They note, as did Perry and Adams (1978), the similarities of
stromatolites to desert rock varnish in that with both, laminations are present and microbes appear to precipitate and
entrap metals and fine particles. They suggested that studies of ocean manganese and stromatolite formation may be
applicable to understanding some aspects of varnish formation.
Recently, 64 suggested a direct biological control of the mineral deposition of fosterite and opal by the fungal phase of
stomatolitic lichens. In another study 65 stated that subaerial biofilms on exposed rocks are accumulations of cell material
and extracellular polymeric substances (EPS) maintaining life in the presence of minimum water, often less than 1%. She
stated that the EPS probably provides a protective layer for cells and prevents cell desiccation with EPS layers and
photoprotective pigments.
7. Conclusions
Until a few years ago, it appeared that a relatively simple model, emphasizing iron and manganese fixation, could
explain the most important aspects of varnish formation. The potential contributions of organic constituents were largely
ignored. In this context, the experimental approach of Krumbein and Jens (1981), Dorn and Oberlander (1981), and the
Palmer, Staley, Adams experiment to attempt to define how organisms interact with metals, is worthwhile. However,
future experiments need to account for other possible aspects of varnish formation in nature, such as light,
polysaccharides, and other humic substances. The role of silicic acid as a cementing, complexing, and emtombing agent
needs to be investigated. Several analyses are suggestive of the presence of silica in varnishes. TOF-SIMS and XPS have
shown Si enriched on outer surfaces and the possible presence of organics compounds. Amino acids and DNA are
present in or on coatings but their presence does not necessarily imply a causative role in varnish formation. Both δ56Fe
and δ13C however, suggest an inorganic mechanism. The presence of organic compounds does not provide proof for a
biological formation mechanism while support for an inorganic mechanism via polymerization of silicic acid seems
plausible. The preservation of labile amino acids, DNA, microbial fossils54 in a hard silica coating could also occur on
Mars.
Acknowledgements
The NSF Integrative Graduate Education and Research Traineeship (IGERT) grant, DGE-9870713, supported this
work in large part. Support was also provided by grants from the Wisconsin and Washington Space Grant Consortium.
The authors wish to thank Mark Englehard, Daniel Gaspar of PNNL, Brian Beard at University of Wisconsin and
especially Jeremy Dodsworth at the University of Washington who was instrumental in the 16S rRNA study , and Henry
Sun, at JPL for ideas and discussions about photochemical processes.References
1. R. S. Perry, M. H. Engel, O. Botta, and J. T. Staley, "Amino acid analyses of desert varnish from the Sonoran
and Mojave Deserts," Geomicrobiology Journal 20(5)(2003).
2. R. I. Dorn, Rock Coatings, Developments in Earth Surface Processes 6 (Elsevier, Amsterdam, 1998), p. 429.
3. B. E. DiGregorio, "Rock varnish as a habitat for extant life on Mars," in Instruments, Methods, and Missions for
Astrobiology IV, R. B. Hoover, G. V. Levin, R. R. Paepe, and A. Y. Rozanov, eds. (SPIE, Bellingham
Washington, 2002), pp. 120-130.
4. B. Hungate, A. Danin, N. B. Pellerin, J. Stemmler, P. Kjellander, J. B. Adams, and J. T. Staley,
"Characterization of manganese-oxidizing (MnII-MnIV) bacteria from Negev Desert rock varnish: implications
in desert varnish formation," Canadian Journal of Microbiology 33, 939-943 (1987).
5. W. E. Krumbein and H. J. Altman, "A new method for the detection and enumeration of manganese oxidizing
and reducing microorganisms," Helgolander wiss. Meeresunters 25, 347-356 (1973).
6. F. E. Palmer, J. T. Staley, R. G. Murray, T. Counsell, and J. B. Adams, "Identification of manganese-oxidizing
bacteria from desert varnish," Geomicrobiology Journal 4(4), 343-360 (1986).
7. S. Taylor-George, F. E. Palmer, J. T. Staley, D. J. Borns, D. J. Curtiss, and J. B. Adams, "Fungi and bacteria
involved in desert varnish formation," Microbial Ecology 9(3), 227-245 (1983).
8. C. G. Engel and R. P. Sharp, "Chemical data on desert varnish," Bull. Geol. Society of America 69, 487-518
(1958).
9. R. M. Potter and G. R. Rossman, "Desert varnish: the importance of clay minerals," Science 196(4297), 1446-
1448 (1977).
10. R. S. Perry and J. B. Adams, "Desert varnish: evidence for cyclic deposition of manganese," Nature 276, 489-
491 (1978).
11. R. I. Dorn and T. M. Oberlander, "Rock varnish," Progress in Physical Geography 6, 317-367 (1982).
12. T. Liu and W. S. Broecker, "How fast does rock varnish grow?," Geology 28(2), 183-186 (2000).
13. H. Bao, G. M. Michalski, and M. H. Thiemens, "Sulfate oxygen-17 anomalies in desert varnishes," Geochimica
et Cosmochimica Acta 65(13), 2029-2036 (2001).
14. J. B. Adams, F. E. Palmer, and J. T. Staley, "Rock weathering in deserts: mobilization and concentration of
ferric iron by microorganisms," Geomicrobiology Journal 10(2), 99-114 (1992).
15. R. I. Dorn and T. M. Oberlander, "Microbial origin of desert varnish," Science 213(4513), 1245-1247 (1981).
16. W. E. Krumbein and K. Jens, "Biogenic rock varnishes of the Negev Desert (Israel): an ecological study of iron
and manganese transformation by cyanobacteria and fungi," Oecologia 50, 25-38 (1981).
17. B. Nagy, L. A. Nagy, M. J. Rigalli, and W. D. Jones, "Rock varnish in the Sonoran Desert: microbiologically
mediated, accumulation of maganiferous sediments," Sedimentology 38, 1153-1171 (1991).
18. R. S. Perry, J. Dodsworth, J. T. Staley, and A. Gillespie, "Molecular analyses of microbial communities in rock
coatings and soils from Death Valley California," Astrobiology 2(4), 539 (2002).
19. K. Sterflinger, W. E. Krumbein, T. Lallau, and J. Rullkotter, "Microbially mediated orange patination of rock
surfaces," Ancient Biomolecules 3, 51-65 (1999).
20. J. T. Staley, J. B. Adams, and F. E. Palmer, "Desert varnish: a biological perspective," in Soil Biochemistry, G.
Stotsky and J.-M. Bollag, eds. (Marcel Dekker Inc., New York, 1992), pp. 173-195.
21. G. Grote and W. E. Krumbein, "Microbial precipitation on manganese by bacteria and fungi from desert rock
and rock varnish," Geomicrobiology Journal 10(1), 49-57 (1992).
22. M. Eppard, W. E. Krumbein, C. Koch, E. Rhiel, J. T. Staley, and E. Stackebrandt, "Morphological,
physiological, and molecular characterization of actinomycetes isolated from dry soil, and monument surfaces,"
Arch. of Micro. 166, 12-22 (1996).
23. C. D. Elvidge, "Distribution and formation of desert varnish in Arizona," M.S. (University of Arozona,
Phoenix, 1979).
24. C. C. Allen, "Desert varnish of the Sonoran Desert-optical and electron probe microanalysis," Journal of
Geology 86(6), 743-752 (1978).
25. C. E. Jones, "Characteristics and origin of rock varnish from the hyperarid coastal deserts of Northern Peru,"
Quaternary Research 35, 116-129 (1991).You can also read
