Chapter 8
Rock varnish

Ronald I. Dorn
Professor of Geography
Geography Department
PO Box 870104
Arizona State University
Tempe AZ 85287-0104
FAX: 480-965-8313


Geochemical Sediments and Landscapes
edited by
David J. Nash (University of Brighton) and Sue J. McLaren (University of Leicester)
Blackwell, London
Chapter 8, p. 246-297

ISBN 1-4051-2159-5 (pdf version)

Nature and general characteristics

Most earth scientists thinking about geochemical sediments envision stratigraphic sequences, not natural rock exposures. Yet, rarely do we see the true coloration and appearance of natural rock faces without some masking by biogeochemical curtains. Geochemical sediments known as rock coatings (Table 8.1) control the hue and chroma of bare-rock landscapes. Tufa and travertine (Chapter 6), beachrock (Chapter 11), and nitrate efflorescences (Chapter 12) exemplify circumstances where geochemical sediments can cover rocks. Perhaps because its ability to alter a landscape’s appearance dramatically  (Fig. 8.1), the literature on rock varnish remains one of the largest in the general arena of rock coatings (Chapter 10 in Dorn 1998).

Table 8.1. Different types of rock coatings (adapted from Dorn, 1998)

Coating Description
Carbonate Skin Coating composed primarily of carbonate, usually calcium carbonate, but sometimes combined with magnesium.
Case Hardening


Addition of cementing agent to rock matrix material; the agent may be manganese, sulfate, carbonate, silica, iron, oxalate, organisms, or anthropogenic.
Dust Film Light powder of clay- and silt-sized particles attached to rough surfaces and in rock fractures.
Heavy Metal Skins Coatings of iron, manganese, copper, zinc, nickel, mercury, lead and other heavy metals on rocks in natural and human-altered settings.
Iron Film Composed primarily of iron oxides or oxyhydroxides 
Lithobiontic Coatings Organic remains form the rock coating, for example lichens, moss, fungi, cyanobacteria, algae.
Nitrate Crust Potassium and calcium nitrate coatings on rocks, often in caves and rock shelters in limestone areas
Oxalate Crust Mostly calcium oxalate and silica with variable concentrations of magnesium, aluminum, potassium, phosphorus, sulfur, barium, and manganese. Often found forming near or with lichens. Usually dark in colour, but can be as light as ivory
Phosphate Skin Various phosphate minerals (e.g. iron phosphates or apatite) that are mixed with clays and sometimes manganese; can be derived from decomposition of bird excrement
Pigment Human-manufactured material placed on rock surfaces by people.
Rock Varnish Clay minerals, Mn and Fe oxides, and minor and trace elements; colour ranges from orange to black in colour produced by variable concentrations of different manganese and iron oxides.
Salt Crust The precipitation of chlorides on rock surfaces
Silica Glaze Usually clear white to orange shiny luster, but can be darker in appearance, composed primarily of amorphous silica and aluminum, but often with iron. 
Sulfate Crust Composed of the superposition of sulfates (e.g., barite, gypsum) on rocks; not gypsum crusts that are sedimentary deposits 

Fig. 8.1 This road cut between Death Valley, California and Las Vegas, Nevada illustrates the ability of 10-100µm-thick rock varnish (darkens debris slope surface on upper half of image) to mask the appearance of the host igneous and metamorphic rocks (light-appearing road cut on lower half of image).

Rock varnish (often called "desert varnish" when seen in drylands) is a paper-thin mixture of about two-thirds clay minerals cemented to the host rock by typically one-fifth manganese and iron oxyhydroxides. Upon examination with secondary and backscattered electron microscopy, the accretionary nature of rock varnish becomes obvious, as does its basic layered texture imposed by clay minerals (Dorn &Oberlander 1982). Manganese enhancement, two orders of magnitude above crustal values, remains the geochemical anomaly of rock varnish and a key to understanding its genesis.

Field observations have resulted in a number of informal classifications. Early field geochemists recognized that varnish on stones in deserts differs from varnish on rock surfaces intermittently flooded (Lucas 1905). Another example of differentiating varnish involves position on a desert pavement clast (Mabbutt 1979): black varnish rests on the upper parts of a pavement clast; a shiny ground-line band varnish at the soil-rock-atmosphere interface; and an orange coating on the underside of cobbles in contact with the soil (Engel & Sharp 1958). Detailed discussions also distinguish different relationships among varnish and the underlying rock (Haberland 1975).

An issue in classification that the introductory reader should be aware of involves the fairly frequent use of the term "desert varnish" or "rock varnish" to describe completely different types of rock-surface alterations (e.g., Howard 2002).  Even experienced desert geomorphologists (e.g., Bull 1984; McFadden et al. 1989) have mistaken surfaces as such basalt ventifacts and geothermal deposits for varnish. Consider just Antarctic research, that has been consistently unable to distinguish between iron films and weathering rinds usually termed either desert varnish (Glasby et al. 1981) or rock varnish (Ishimaru & Yoshikawa 2000) from manganese-rich coatings (Glazovskaya 1958; Dorn et al. 1992).

 The problem for any student new to the topic of manganese-rich varnish is that the vast majority of papers are written by investigators who collect only a few samples from a few locales and simply assume that their samples somehow, magically, are equivalent to varnishes collected in completely different biogeochemical settings.  The only formal classification of rock varnish thus far presented (Dorn, 1998, p. 214-224) maintains the goal of forcing varnish researchers to ensure that they have sampled the same variant in any comparisons. For the sake of clarity and brevity, this chapter presents illustrations only for varnishes formed in subaerial settings.

Distribution, field occurrence and geomorphological relations

Rock varnish grows on mineral surfaces exposed by erosional processes, either in weathering-limited landscapes like deserts (e.g., White 1993) or in settings dominated by rapid erosion such as glaciated terrain (e.g., Whalley et al. 1990). Thus, one first-order control on distribution rests with geomorphic processes exposing rock faces.  Another first-order control on distribution rests in whether the varnish formed in a geochemical landscape (Perel'man 1961) favorable for stability; exposure to minimal acidity helps maintain the manganese and iron in their oxidized and immobile condition in varnish.  A second-order issue turns on the nature of the erosional process.  Spalling often exposes planar joint faces that already have rock coatings developed in subsurface settings (Coudé-Gaussen et al. 1984; Villa et al. 1995; Dorn 1998).

 Rock varnish exists in geomorphic catenas (Haberland 1975; Palmer 2002), much in the way that soils vary along slopes (Jenny 1941). Taking a simple basalt hill in the western Mojave Desert of California, Palmer (2002) mapped the abundance of black rock varnish, orange iron films, and surfaces experiencing rapid-enough erosion that they lacked field-observable rock coatings (Fig. 8.2). Rock varnish occurs more frequently in on metre-sized colluvium at the hill crest. In contrast, iron films, that initially formed on clasts formerly buried in soil, increasingly dominate on subaerial slope positions experiencing greater soil erosion from grazing (Palmer 2002) or Holocene erosion (Hunt and Wu, 2004).

Fig. 8.2.  Rock varnish varies considerably over short distances, over a single boulder and over a single hillslope. The upper part of this figure illustrates that varnish thickness varies considerably over a single granodiorite boulder on the Tioga-3 moraine at Bishop Creek, about 19±1  ka 36Cl years old (Phillips et al. 1996).  The thickest varnishes grow in shallow "dish shaped" microbasins.  Deeper rock depressions typically host varnishes experiencing erosion from acid-generating fungi and cyanobacteria.  The lower portion of this figure presents a rock coating catena (adapted from Palmer 2002) for three rock coating types, shows variability on a basaltic hillslope in the Mojave Desert, California. Note that rock varnish dominates the more stable hill crest; iron films increase in abundance towards the slope bottom, while rock surfaces mostly lacking either coating occurs in a similar abundance on all slope positions.

 Palmer's (2002) general observation that different geomorphic positions influence rock varnish carries over to a single boulder. Just considering varnish thickness, millimetre-scale topographic highs host thinner varnishes than "microbasins" or broad depressions a few millimetres across (Fig. 8.2).  Depressions more than a few millimetres deep favor retaining enough moisture to host varnish-destroying microcolonial fungi (Fig. 8.3), cyanobacteria, or even lichens. Granodiorite boulders on a moraine in the Karakoram Mountains, for example, illustrate the importance of aspect where varnish grows on south-facing surfaces, while lichens and salt efflorescence occupy north-facing surfaces (Waragai 1998).

Fig. 8.3 Originally named black globular units (Borns et al. 1980) for their appearance under a hand lens, microcolonial fungi are common inhabitants on desert rocks (Staley et al. 1982) that experience warm season convective precipitation.  Originally thought to be agents of varnish (Staley et al. 1983) formation, the lack of manganese-enhancement by these and other fungi (Dorn & Oberlander 1982) and their ability to dissolve varnish (Dorn 1994) makes these organisms either adventitious or destructive agents, but not formative (Dragovich 1993).

 The varnish literature is filled with misunderstandings surrounding its distribution, field occurrence and geomorphic relations. Consider one of the most common errors, equating darkness in a field sample with age, a generalization derived from observing the darkening of alluvial-fan units over time.  If geomorphic units darken over time, would not individual hand samples? This is not necessarily the case. The darkest varnishes are usually those that grow first in rock crevices (Coudé-Gaussen et al. 1984; Villa et al. 1995) and are then "born dark" as spalling exposes fracture varnish. Rather than detail each misunderstanding here, Table 8.2 summarizes these peculiar components of the literature.

Table 8.2 A few of the literature misunderstandings surrounding rock varnish and its environmental relations. See chapter 10 in Dorn (1998) for a more detailed accounting of the history of thought on varnish.

Topic Misunderstanding Resolution
Accretion or rock weathering A century of debate exists on whether varnish accretes on (Potter and Rossman 1977) or derives from the underlying rock (Hobbs 1918) Low (Dorn and Oberlander 1982) and high resolution (Krinsley 1998) electron microscopy and trace element geochemistry (Thiagaran and Lee 2004) fails to show evidence of anything other than an accretionary process
Adventitious organisms Varnishes hosts a wide array of organisms (Kuhlman et al. 2005; Schelble et al. 2005;  such as lichens (Laudermilk 1931), microcolonial fungi (Borns et al. 1980; Staley et al. 1983; Gorbushina et al. 2002) gram-positive bacteria (Perry et al. 2003), cyanobacteria and other organisms (Krumbein 1969). The simple presence of organics in association with varnish bears little importance to varnish genesis, unless biotic remains explicitly link to varnish-formation processes. Spatially precise geochemical analyses reveal the difference between adventitious organisms (Dragovich 1993) and those bacteria that enhance Mn and Fe (Dorn & Oberlander 1992; 1998).
Case hardening Rock varnish is too thin to be an effective case hardening agent. Cation leaching (Dorn and Krinsley 1991) from varnish provides materials for redeposition within the weathering rind — thus case hardening the outer shell of rock (Haberland 1975).
Darkness shows time Darker appearing surfaces are older than younger appearing surfaces (McFadden et al. 1989; Reneau 1993) This misunderstanding causes a number of errors in geomorphic analyses. Eleven factors other than time influence darkness (Table 10.7 in Dorn, 1998)
Desert restriction Desert varnish occurs in deserts and not elsewhere, and this distribution somehow relates to mode of formation. Rock varnish occurs in every terrestrial weathering environment (Dorn and Oberlander 1982; Douglas et al. 1994; Lucas 1905; Whalley et al. 1990). More frequent occurrence in deserts relates greater landscape geochemical stability (Perel'man 1961) for Mn and Fe cements.
Lithology Preference Rock varnish grows best on certain lithologies such as fine grained extrusive rocks. Rock varnish forms on all lithologies whose surface stability exceeds rates of varnish formation. 
Lustre The shiny appearance of varnish relates to age or wind abrasion. Neither is true. Some sheen derives from very smooth and Mn-rich varnishes. The shiniest varnishes, found at the ground line on desert pavement cobbles, are only a few microns thick and intercalate with silica glaze.
Manganese and iron abundance When averaged over an area, the total accumulation of manganese and iron increases over time. Major problems confound this naïve belief, including exposure of manganese-rich fracture coatings (Douglas et al. 1994), erosion of faster-forming varnishes (Dorn & Oberlander 1982), and role of microenvironments (Whalley et al. 1990).
Mars analog Assuming that varnish is a biogenic rock coating, the occurrence of Mn-rich varnishes on Mars provides evidence of life or former life. The use of varnish as a bioindicator of Martian life is untenable until abiotic origins of Mn-enhancement are falsified for Martian conditions.
Surface stability Since cosmogenic nuclide analyses show erosion of rock surfaces (Nishiizumi et al. 1993), and geomorphic surfaces hosting varnish such as desert pavements experience instability (Peterson 1981), how can varnish exist for 105 years (Liu and Broecker 2001; Liu 2003)?  Rock surface erosion rarely occurs in a linear fashion. Larger clasts can host zero erosion rates, while adjacent clasts experience much more rapid erosion. Linear rates of boulder erosion remains unusual. Similarly, landforms can exhibit signs of instability such as erosion of fines, even though particular rock faces can remain stable. 
Thickness  Thickness increases over time The thickest varnishes grow in wetter micropositions that foster erosion of the underlying weathering rind or on spalled fracture faces. Thus, although varnish does thicken over time if all other formation factors are held constant, the thinnest varnishes can be the oldest.


Macro- and micromorphological characteristics

Regions with abundant vegetative cover host rock varnish on bare-rock surfaces that are too small in extent to alter signatures in remotely-sensed imagery.  That is why arid regions host a large number of remote sensing investigations about the macromorphological characteristics of varnish (Kenea 2001).  Alluvial fans are favorite study sites (White 1993; Farr & Chadwick 1996; Milana 2000), in part because of the dramatic colour changes produced by progressive increases in the aerial extent of varnishing (Fig. 8.4) and increasing darkness as fracture varnishes are exposed by boulder spalling (Villa et al. 1995).

Fig. 8.4 Rock varnish on Hanaupah Canyon alluvial fan, Death Valley, viewed at different scales.
         Top Row: SPOT satellite image where darkness first increases and then decreases with age.  As varnish coverage increases, older sections of the Bar and Channel segment darken — reaching maximum darkness on the smooth pavement (ground view upper right) segment.  Channel incision starts to lighten appearance as varnished cobbles erode into gullies, with lightening increasing as channel incision progresses to the point where rounded ballenas expose a mix of calcrete and varnished cobbles.
         Bottom Row: Electron microsope imagery all revealing layering patterns in rock varnish, collected from Hanaupah Canyon alluvial fan, Death Valley.  LEFT: BSE electron image reveals average atomic number, where the bright layers are rich in manganese and iron and the darker layers are richer in clay minerals.  MIDDLE: SE image shows shape, with the arrows identifying the contact between the rock in the lower right and the layered varnish.  RIGHT: HRTEM image is at a much greater magnification, showing that there are layers within layers, where the layered nature of clay minerals (Potter & Rossman 1977) imposes the basic texture.

 Regional climatic variations alter the micromorphological character of rock varnishes, where abundance of aeolian dust appears to be a key factor. The vast majority of rock varnishes have micromorphologies that range between lamellate and botryoidal (Fig. 8.5).  A paucity of dust allows a clustering of varnish accretion, thus producing botryoidal forms.  An abundance of dust smothers the tendency of Mn-accumulation around nucleation centers, producing lamellate micromorphologies. A variety of other factors, however, can influence micromorphology including: clay mineralogy; manganese abundance; epilithic organisms such as microcolonial fungi; morphology of the underlying rock; varnish thickness; aeolian abrasion; proximity to soil; and other microenvironmental factors (Dorn 1986).

Fig. 8.5 Varnish micromorphology forms ranges from botryoidal (top row) to lamellate (bottom row). The top row shows two types of imagery of botryoidal varnishes from Kitt Peak, Arizona. The view on the left shows the topography by secondary electrons. The view on the right shows these structures from the bottom upwards with backscattered electrons, showing the layering structures inside each nucleation center. The bottom row presents SE images of clay minerals accreting on rock varnish in Death Valley, California. Note, in the lower left image, how the individual clay platelets overlap as they cement onto the surface. The lower right image shows how the clays impose a lamellate structure in cross-section, as first noticed by Potter and Rossman (1977).

 The dominant cross-sectional texture of rock varnish is layering (Fig. 8.4, bottom row). Even the highest resolution transmission electron microscopy (HRTEM) views of varnish consistently show layering textures (Krinsley et al. 1995) imposed by its clay mineralogy (lower right of Fig. 8.4).  Even when the surface morphology is botryoidal, cross-sectional texture remains layered internally within each botryoidal mound (upper right of Fig. 8.5).  All meaningful explanations of varnish genesis must explain varnish textures as well as varnish chemistry.

Chemistry and mineralogy

 Analytical chemist Celeste Engel completed the first M.S. thesis research on rock varnish and teamed up with geomorphologist Robert Sharp to write a seminal paper on rock varnish in the Mojave Desert (Engel & Sharp 1958). The major elements are O, H, Si, Al, and Fe in approximately equal abundance with Mn. The key mystery of varnish formation (von Humboldt 1812; Lucas 1905; Engel & Sharp 1958; Jones, 1991) is how to explain the great enrichment in Mn, normally a trace element in soils and rocks. Varnish Mn:Fe ratios vary from less than 1:1 to over 50:1, in contrast with Mn:Fe ratios of about 1:60 in the earth's crust. Early microprobe analyses showed measurements higher than 15% Mn  (Hooke et al. 1969). Concentrations of over 80% Mn in focused spots occur on budding bacteria forms (Dorn 1998).

 Mn (II) is the most important soluble form, generally stable in the pH range of 6 to 9 in natural waters (Morgan & Stumm 1965). Since Mn (III) is thermodynamically unstable in a desert environment lacking abundant humic acids, Mn (IV) is the primary insoluble oxyhydroxide found in varnish with average valencies of about +3.8 to +3.9  (Potter & Rossman 1979a; McKeown & Post 2001).  The general Mn (II) oxidation reaction is:

Mn2+ + 1/2 02 + H2O —> MnO2 + 2H+                                                                               (1)

for an oxidation state state of (IV).

The inherent heterogeneity of the varnish became abundantly clear in analyses of minor elements. In order of decreasing concentration, these tend to be: Ca, Mg, Na, K, Ti, and P. Like Mn, minor elements vary considerably (0.5 - 1.5%) with Ba, and Sr as the most abundant trace elements, along with Cu, Ni, Zr, Pb, V, Co, La, Y, B, Cr, Sc, and Yb found in all varnishes in order of decreasing concentration (Engel & Sharp 1958). Cd, W, Ag, Nb, Sn, Ga, Mo, and Zn were found in some, but not all varnishes (Lakin et al. 1963). With the acquisition of data from newer techniques using electron microprobes (Liu 2003), neutron activation (Bard 1979), PIXE (Dorn et al. 1990), inductively coupled plasma mass spectrometry (Thiagarajan & Lee 2004), and ion microprobes (O'Hara et al. 1989), heterogeneity remains the rule of chemical composition at all spatial scales: from micron to micron within a single depth profile or chemical transect, between locations on the same rock, between different rocks from the same area and between different areas (Table 8.3).  [The section of this paper discussing the palaeoenvironmental significance of varnish focuses on variations in varnish chemistry with depth.]

Table 8.3. Examples of elemental variation exhibited in bulk chemical analyses of rock varnishes found in desert regions. Results are normalized to 100% with measurements by PIXE , with na indicating not available and bld below the limit of detection.

Trail Fan, Death Valley
Manix Lake, Mojave Desert
Makanaka Till, Hawaii
Sinai Peninsula, Egypt
Petroglyph South Australia
Ingenio, Peru Desert
Ayers Rock, Australia
Former Rock Fracture
> 1m Above Soil
With Silica Skin
>1m Above Soil
> 1 m Above Soil
At Soil Surface
From Rock Fracture

The following generalizations remain unchanged from the original observations of Engel and Sharp (1958): 1) silica and aluminum, taken together, comprise the bulk of rock varnish — consistent with clays being the dominant mineralogy; 2) manganese and iron oxides typically comprise one-quarter to one-third of rock varnish with high point-to-point variability at scales from nanometres to kilometres; 3) minor elements display variable patterns. Mg, K, and Ca are correlated with clays in cation-exchange positions and not with manganese. Barium often correlates with S in barium sulfates (O'Hara et al. 1989; Dorn et al. 1990; Cremaschi 1996). In other cases, Ba correlates with Mn (Liu 2003). Ti can correlate with Fe in titanomagnetite detrital grains (Mancinelli et al. 2002), but more often Ti is not well correlated with any major element (Dorn 1998); and 4) trace heavy metals generally correlate with Mn abundance, and sometimes with Fe, due to the scavenging properties of these oxyhydroxides (Jenne, 1968; Thiagarajan and Lee, 2004; Tebo et al., 2004).

 Varnish minerals were originally reported to be amorphous (Engel & Sharp 1958), with goethite (Scheffer et al. 1963) and ferric chamosite (Washburn 1969) as important components.  Seminal research conducted with infrared spectroscopy, X-ray diffraction and electron microscopy at the California Institute of Technology revealed that the bulk of rock varnish is composed of clay minerals (Potter & Rossman 1977), dominantly illite, montmorillonite, and mixed-layer illite-montmorillonite. As previously noted, the layering seen in varnish at all scales reflects this clay mineralogy (Figs. 8.4 and 8.6). Subsequent research has confirmed the dominance of clay minerals (Diaz et al., 2002; Dorn, 1998; Dorn and Oberlander, 1982; Israel et al., 1997; Krinsley, 1998; Krinsley et al., 1990; Krinsley et al., 1995; Potter and Rossman, 1977, 1979c; Probst et al., 2002), although there is some disagreement (Perry et al., 2006).

Fig. 8.6  HRTEM imagery reveals manganese and iron minerals that appear to be moving from the granular remnants of bacterial sheaths into adjacent clay minerals. The granular textures found between layered varnish consist of Mn-Fe precipitates on bacterial remains (Krinsley 1998).  Images derive from (A) Nasca, Peru (B,E) Death Valley (C,D) Kaho'olawe, Hawai'i, and (F) Antarctica varnishes with scale bars in nanometres.

 Clay minerals are cemented to the underlying rock by oxides of manganese and iron. Birnessite ([Na,Ca]Mn7014 • 2.8H20) and birnessite-family minerals are the dominant manganese minerals in black varnish, and hematite is a major iron oxide in both black varnish and orange iron films (Potter & Rossman 1977,  1978, 1979a, 1979b, 1979c; McKeown & Post 2001). Birnessite-family minerals are found frequently in microbial deposits (Tebo & He 1998; Tebo et al., 2005). McKeown & Post's (2001: 712) summary perspective holds that:

"[e]ven if analysis methods are improved, the situation will remain complicated by the flexibility and great variety of Mn oxide structures. The common elements of these structures enable them to easily intergrow with an transform with one another. Furthermore, many of the phases, particularly the layered structures, readily exchange interlayer cations in response to even slight changes in chemistry on a microscale". The HRTEM imagery (Krinsley et al. 1995; Krinsley 1998) reveals that the iron and manganese oxides that cements clay minerals together exist in this layered structure at the nanometre scale, derived from what appears to be sheaths of bacteria (Fig. 8.7).  The granular texture of the bacteria sheaths can be seen releasing nanometre-scale Mn-Fe that could be unit cells (Fig. 8.6)  into the weathered edges of the clay minerals (Fig 8.8) (Dorn 1998).  The ability to undergo incomplete transformation between phases at the nanometre scale would certainly be consistent with a disordered todorokite/birnessite-like phase (cf. Potter & Rossman 1979a; McKeown & Post 2001).

Fig. 8.7. Budding bacteria morphologies actively concentrate manganese from Crooked Creek, Kentucky (top), Negev Desert (middle) and Antarctica (bottom), as seen by energy dispersive "spots" focused on the bacteria, composed with defocused analyses on adjacent uncolonized varnish. The Negev Desert hyphae resemble Metallogenium in morphology (Perfil'ev et al. 1965). The squiggly arrow in the Antarctica (bottom) frame shows hyphae extending from bacterial-sized cells, exhibiting a Pedomicrobium budding bacteria morphology.

 Detritus, both organic and inorganic, often settles in morphological depressions on varnish.  These pieces are sometimes trapped by accreting rock varnish, leading to a wide variety of other minerals that are sometimes found in varnish samples such as quartz, feldsaprs and magnetite (Potter & Rossman 1979a; Dorn 1998: 198; Mancinelli et al. 2002).

Fig. 8.8 Clay minerals appear to be weathering by the insertion of Mn-Fe in samples from Antarctica (A,B), Peru (C), and Death Valley (D). Nanometre-scale cells of Mn-Fe appear within detrital clay mineral grains (arrows  in A and B), essentially weathering 001 planes of exfoliation (Robert & Tessier 1992). In many cases, the feathering is associated with granular textures (cf. Fig. 8.6). At higher magnifications (C,D), disorganized wavy layers from clay weathering rest next to regularly spaced lattice fringes (with spacing that is consistent with illite, smectite, chlorite, and interstratified clay—textural interstratification) (Robert et al. 1990). Scale bars in nanometres.

Mechanisms of formation or accumulation

The genesis of most varnish constituents are explained easily.  The clay mineral building blocks of varnish are ubiquitous in terrestrial weathering environments where rock coatings are found (Dorn 1998) and hence require simple dust deposition by aeolian processes.  Minor and trace elements come adsorbed to clays, with subsequent enhancement by scavenging properties of Mn-Fe oxyhydroxides (Jenne 1968; Forbes et al. 1976; Thiagarajan and Lee 2004).

 Explaining the great manganese enhancement in varnish (von Humboldt 1812; Lucas 1905; Engel & Sharp 1958) has long been recognized as a key to understanding varnish formation. How clay minerals and manganese selectively interact, however, is just as important a key in discriminating relevant formative processes as concentrating Mn (Potter & Rossman 1977).  The mutual dependency of clay minerals and manganese can be realized by understanding what happens when one is lacking.  Dust films, silica glaze and other rock coatings may result from clay mineral accretion without Mn (Dorn 1998). Conversely, a different type of rock coating, heavy-metal skins, accrete when manganese precipitation occurs without clay minerals (see chapter 8 in Dorn 1998).  Desert varnish accumulates only when and where nanometre-scale fragments of manganese, as seen in HRTEM imagery, cements broken and decayed fragments of clay minerals to rock surfaces.   Viable models of rock varnish formation must turn on processes by which manganese enhancement takes place and processes by which manganese and iron interact with clay minerals.

Four general conceptual models have been proposed to explain varnish formation (Fig. 8.9). This section first presents the four proposed hypotheses of varnish genesis and then evaluates strengths and weaknesses.

Figure 8.9: [click here to see higher resolution] Conceptual models of rock varnish formation. (A) Abiotic enhancement by iterations of acid solutions separating Mn2+  followed by oxidizing conditions concentrates Mn in varnish (Hooke et al., 1969; Smith and Whalley, 1988). (B) Polygenetic clay-bacteria-interactions where varnish formation starts with the oxidation and concentration of Mn (and Fe) by bacteria.  Wetting events dissolve nanometre-scale fragments of Mn.  Ubiquitous desert dust supplies interstratified clay minerals. The Mn-bacterial fragments fit into the weathered edges of clays, cementing clays much like mortar cements brick.  (C) Lithobionts or their organic remains (e.g. spores, polysaccharides and other humic substances) play a role in binding varnish and concentrating Mn and Fe oxides and oxyhydroxides. (D) Silica binding of detrital grains, organics, and aerosols (Perry and Kolb, 2003; Perry et al., 2006).

Abiotic hypothesis of varnish genesis

Proponents of the abiotic hypothesis purport that varnish forms without the aid of biological enhancement of Mn (Linck, 1901; Engel and Sharp, 1958; Hooke et al., 1969; Moore and Elvidge, 1982; Smith and Whalley, 1988). To increase Mn:Fe ratios two to three orders of magnitude above crustal values, abiotic processes rely on greater mobility of divalent manganese over ferrous iron where small pH fluctuations dissolve Mn but not Fe (Krauskopf, 1957). Mn released by slightly acidic precipitation is then fixed in clays after water evaporation or change in pH.

Redox reactions between Mn (II) and Mn (IV) are largely governed by pH.  In natural waters, Mn (II) oxidation requires pH values greater than 8.5 to oxidize homogeneously within weeks to months without microbial assistance (Morgan and Stumm, 1965). Laboratory experiments reveal that abiotic oxidation from Mn (II) to Mn (IV) requires two steps (Hem and Lind, 1983); Mn first precipitates as an oxyhydroxide (MnOOH) that secondly forms tetravalent Mn oxides (MnO2).  Abiotic oxidation of Mn (II) may be increased by the presence of abundant Fe oxyhydroxides found in varnish  (Junta and Hochella, 1994), but Fe-catalyzed oxidation results in Mn (III)-bearing oxyhydroxides (Junta and Hochella, 1994) that are not found in varnish samples (cf. Potter and Rossman, 1979c; McKeown and Post, 2001).

In contrast, microbial oxidation of Mn (II) to Mn (IV) is quite rapid at the near neutral pH values (Tebo et al., 1997; Tebo et al., 2004) found on rock varnish (Dorn, 1998). The process appears to be extracellular where Mn oxides create casts on cell walls (Krinsley, 1998). There are both gram-positive and gram-negative bacteria that form a diverse phylogenetic group of sheathed and appendaged bacteria capable of oxidizing Mn (e.g. Perfil'ev et al., 1965; Khak-mun, 1966; Dorn and Oberlander, 1981, 1982; Peck, 1986; Hungate et al., 1987). Environmental and laboratory studies suggest that the microbial oxidation of Mn (II) results in the formation of Mn (IV) without lower Mn (III) valences (Tebo and He, 1998; Tebo et al., 2004).

The geography of varnish poses critical obstacles, making purely abiotic formation of varnish difficult to explain. Varnishes found in wetter climates are typically higher in manganese than those in arid, more alkaline settings (Dorn, 1990) - a result that would not be predicted by the high pH requirements of abiotic oxidation.  Black, Mn-rich varnishes occur in a host of moist environments such as within joint faces in glacial Norway (Whalley et al., 1990) and Iceland (Douglas et al., 1994).  In contrast, hyperarid deserts are dominated by varnishes with a relative paucity of manganese; orange to burgundy-colored varnishes cover rocks in hyperarid deserts, such as the Atacama, and locally alkaline settings such as those adjacent to deflating salt playas in western North America (Dorn, 1998).  Repeated oscillations in pH are the key to abiotic concentration of manganese over iron (cf. Engel and Sharp, 1958; Hooke et al., 1969; Smith and Whalley, 1988).  However, the very environments where one would expect the least number of these pH fluctuations are the humid environments where Mn-rich varnishes are most common (Dorn, 1990, 1998) - providing a geographic conundrum for the abiotic enhancement mechanism.

Other geographical problems with the abiotic hypothesis (Dorn, 1998, pp. 242-243)  include: rapid pH fluctuations in the range to mobilize and oxidize Mn where varnish does not commonly occur, in such as places as the rainshadows of Mauna Loa and Mauna Kea on Hawai’i; purely inorganic Mn fixation experiments yielding too low concentrations of Mn; millimetre-scale distribution of varnish mimicking the pattern of microbial colonization; an inability to explain botryoidal varnish textures by abiotic processes; and an inability to produce varnish experimentally without microorganisms.

One of the most serious problems with the abiotic explanation rests in a simple analysis of rates of varnish formation.  Based on detailed analyses of over ten thousand microbasins (Liu and Broecker, 2000; Liu, 2003; Liu 2006; Liu and Broecker, 2006), consistent with independent observations (Dorn, 1998), rates of varnish formation are on the order of a few microns per thousand years. If Mn is enhanced abiotically, there would be no reason for such a slow rate of varnish formation.  Multiple dust deposition and carbonic acid wetting iterations take place annually, even in drought years.  Under this model, a bit of Mn would be leached from dust with each event and in turn mix with other constituents to accrete in varnish layers.  Such progressive leaching would be relentless without a rate-limiting step, and varnish accretion would be from 102 to 104 times faster than real varnishes, that is if abiotic enhancement led to varnish formation.

Although arguments could potentially be made against each of these problems with an abiotic explanation of Mn-enhancement, such accommodations have not yet been forthcoming in the literature. Abiotic processes are certainly involved in varnish formation (cf. Bao et al., 2001), including trace element enhancement with wetting (Thiagarajan and Lee, 2004) and the process of clay cementation (Potter and Rossman, 1977; Potter, 1979: 174-175; Dorn, 1998). However, these processes alone would not generate a rock varnish.

Lithobionts or their organic remains produce and bind varnish

The general biological argument includes of a number of researchers who report observing microbial remains growing on varnish (e.g., Allen et al. 2004), or report organic compounds such as bacterial DNA (e.g., Perry et al. 2002) within varnish. Lichens (Laudermilk 1931; Krumbein 1971), cyanobacteria (Scheffer et al. 1963; Krumbein 1969), microcolonial fungi (see Fig. 8.3; cf. Staley et al. 1982; Allen et al. 2004: 25), pollen (White 1924), peptides (Linck 1928), refractory organic fragments (Staley et al. 1991), gram-negative non-sporulating cocci bacteria (Sterflinger et al. 1999), and amino acids from gram-positive chemo-organotrophic bacteria (Warsheid 1990; Nagy et al. 1991; Perry et al. 2003) illustrate only spatial correlation with varnish. [See Dorn (1998, p. 238-9, 243) for a more complete list of references.] Fatty acid methyl esters analyzed from Mojave Desert varnish do not provide any evidence that adventitious gram-positive bacteria responsible for these fatty acids play a role in varnish formation (Schelble et al., 2005). A great leap forward in the study of microbes potentially involved in varnish genesis, however, rests in understanding phylogenetic affiliations and in estimating the total number of microbes actually present— in this case in fracture varnish collected from desert pavement cobbles of the Whipple Mountains of eastern California after a winter rain (Kuhlman et al., 2005, 2006).  About 108 cells of bacteria per gram dry weight of varnish summarizes the story: varnish represents a habitat occupied by a plethora of bacteria, with a significant presence of gram-negative Proteobacteria groups and Actinobacteria (Kuhlman et al., 2005, 2006).  Spatial contiguity of organisms and organic remains with rock varnish, however, is simply insufficient to explain great Mn-enhancement or clay cementation.

The aforementioned growing number of observations of apparently adventitious organisms and organic matter suffer from a core problem: there is no process by which the organisms or their remains produce rock varnish. The most cited of these purported agents — microcolonial fungi  (see Fig. 8.3 and Borns et al. 1980; Staley et al. 1983; Gorbushina et al. 2002) — either dissolve or co-exist with varnish (Dragovich 1993; Dorn 1998).  Even research on cultured Mn-oxidizing bacteria (Krumbein 1969; Dorn & Oberlander 1981; Krumbein & Jens 1981; Dorn & Oberlander 1982; Palmer et al. 1985; Hungate et al. 1987) often miss a key criteria  — morphologically-similar shapes actually seen enhancing Mn in situ  (Fig. 8.7)  (cf. Perfil'ev et al. 1965; Khak-mun 1966; Dorn & Oberlander 1981; Dorn & Oberlander 1982; Peck 1986). Research into the microbial habitat of rock varnish is just starting, and the future rests in following in the footsteps of Kuhlman et al. (2005, 2006) as long as this research is matched with in situ studies.

 Rates of varnish formation in desert environments should make readers very skeptical of the growing plethora of biological agents purported to be involved in varnishing.  In places where varnishing rates have been measured in warm desert environments, typically one to twenty bacterial diameters accumulate each thousand years (Liu 1994; Liu & Broecker 2000).  Rates would be several orders of magnitude higher if all of the observed rock-surface organisms, extracted organic matter associated with varnish and rock surfaces, and cultured biological agents actually play a roll in making desert varnish.  Consider the types of amino acids extracted from varnishes; the apparent lack of D-alloisoleucine and concentrations of serine suggest that the extracted amino acids are only centuries old — not the thousands of years that would indicate a formative role. If all of the biologocal agents suspected of varnish formation are truly involved,  we would be able to see desert rock surfaces encrusted at rates such as seen in travertine contexts (cf. Chafetz et al. 1998). Furthermore, if these commonly seen organisms made varnish, cross-sections would look like biotically-generated coatings where fossilized remains are common (Francis 1921; Ferris et al. 1986; Konhauser et al. 1994; Chafetz et al. 1998; Kennedy et al. 2003).

 Fossilized casts of Mn-enhanced bacterial remains are rare in desert varnish; the intra-varnish remains seen in HRTEM and BSE imagery often resemble gram-negative budding or appendaged bacteria in their morphology (Dorn & Meek 1995; Krinsley 1998; Dorn 1998).  Simply finding organic remains such as amino acids, DNA or even culturing Mn-oxidizers obfuscates the key question: what is the evidence that certain microbes supply the Mn (and Fe) that cement clays to rock surfaces?  Researchers wishing to advance any single adventitious agent to the role of formative agents must present clear evidence of in situ  textural connections to formative processes involving Mn Enhancement.

Silica binding model

The most recent conceptual model proposed to explain varnish formation focuses on silica as the key agent of varnish formation (Perry and Kolb, 2003; Perry et al., 2006).  The hypothesis starts with dissolution of silica from anhydrous and hydrous minerals.  The next step has the silica gelling, condensing, and hardening.  A romantic article title "baking black opal in the desert sun" (Perry et al., 2006) emphasizes the importance of the hot desert environment to this model.  In the process of silica binding, detrital minerals, organics, and dust from the local environment are bound into varnish.  There is no ambiguity about the key role of silica in the envisioned process: "[s]ilicic acid and (di)silicic acid and polymers form complexes with ions and organic molecules, especially those enriched in hydroxy amino acids ... The centrality of silica to this mechanism for cementing and forming coatings means that silica glazes and desert varnish should be considered part of the same class (Perry et al., 2006: 520)."

There is no question that the importance of amorphous hydrated silica (opal) and silica minerals (e.g. mogonite) have not been emphasized in varnish research. Furthermore, the process of silica binding as envisioned (Perry and Kolb, 2003; Perry et al., 2006) would be compatible with other models of varnish formation.  In other words, the movement of water that remobilizes varnish constituents (Dorn and Krinsley, 1991; Krinsley, 1998) would be compatible with and could explain observations of silica (Perry and Kolb, 2003; Perry et al., 2006).

There are, however, multiple fatal flaws with silica binding as the key process in varnish formation. First, the silica binding process does not explain Mn-enhancement seen in varnish or the Mn-minerals seen in varnish.  Second, although silica dissolution is slow on rock surfaces, it is not a rate-limiting step capable of slowing varnish accretion to the scale of a few microns per millennia.  The growth rate of silica glaze within decades (Curtiss et al., 1985; Gordon and Dorn, 2005), perhaps from silica binding (Perry and Kolb, 2003; Perry et al., 2006) or perhaps from another process (chapter 13 in Dorn, 1998), indicates that this model cannot explain varnish accretion. Third, the silica binding model rides on the opinion that clay minerals are not a dominant part of rock varnish (Perry et al., 2006), a view incompatible with data sets several orders of magnitude more extensive (Diaz et al., 2002; Dorn, 1998; Dorn and Oberlander, 1982; Israel et al., 1997; Krinsley, 1998; Krinsley et al., 1990; Krinsley et al., 1995; Liu, 2003; Liu and Broecker, 2006; Potter and Rossman, 1977, 1979c; Probst et al., 2002).  Fourth, this model does not accommodate the geography of rock coatings.  For example, why would silica glazes dominate on Hawaiian basalt flows (Curtiss et al., 1985; Gordon and Dorn, 2005) and not manganiferous rock varnish? Abundant examples of this juxtaposition exist (e.g. Dorn, 1998), and one of the closest Mars analog exists on the Tibetan Plateau, where rock varnish rests adjacent to silica glaze and other rock coatings (Dorn, 1998: 367-371).  The silica binding model cannot explain this geographical pattern. Fifth, "baking" is a requirement of the silica binding model, but varnish occurs in subsurface and cold-climate settings where the requisite heat and light do not exist (Anderson and Sollid, 1971; Dorn, 1998; Dorn and Dragovich, 1990; Dorn et al., 1992; Douglas, 1987; Douglas et al., 1994; Villa et al., 1995; Whalley et al., 1990). Sixth, the silica binding process does not explain the characteristics of the varnish microlaminations data set discussed in a later section.

Polygenetic model

The polygenetic model of rock varnish formation combines biotic enhancement of Mn with abiotic processes (Dorn 1998). It builds on the mineralogy research of Potter and Rossman (1979a, 1979b, 1979c) and studies of bacteria fossil remains within varnish layers (Dorn & Meek 1995; Krinsley et al. 1995; Dorn 1998; Krinsley 1998). In brief, weathered remains of Mn-rich bacterial casts (Fig. 8.6) cement weathered clay minerals to rock surfaces.

 Varnish formation probably starts with ubiquitous dust deposition on rock surfaces; budding bacteria (Dorn & Oberlander 1982) actively concentrate Mn and Fe (Fig. 8.7). Some of these bacterial "sheaths" become microfossils (Dorn & Meek 1995; Dorn 1998; Krinsley 1998) that then decay (Fig. 8.6).  Mn-enhancement by other microbes, such as gram-positive bacteria, could certainly contribute; but in situ  evidence of Mn-enhancement and preservation of Mn-Fe "bacterial casts" has heretofore been lacking for morphologies other than budding or appendaged bacterial forms.  The exact mechanism for oxidation of Mn2+ is uncertain, perhaps involving heteropolysaccharides containing the protein that catalyzes the oxidation reaction with molecular oxygen. Mechanisms for Mn2+ binding and oxidation, however, have yet to be demonstrated for varnish.

 Decay of Mn-Fe-encrusted bacterial casts mobilize nanometre-scale Mn and Fe then cement clay minerals (Fig. 8.8).  At the highest HRTEM magnifications, irregular layers of Mn-Fe-cemented clays rest next to regularly spaced lattice fringes with spacing that is consistent with illite, smectite, chlorite, and interstratified clay-textural interstratification.  [The reader should note that this discussion is at the nanometre-scale and that the interlarying of different types of rock coatings discussed elsewhere is at the much coarser micrometre scale.] Potter (1979: 174-175) argued for this step without the benefit of supporting HRTEM imagery:

"Deposition of the manganese and iron oxides within the clay matrix might then cement the clay layer...the hexagonal arrangement of the oxygens in either the tetrahedral or octahedral layers of the clay minerals could form a suitable template for crystallization of the layered structures of birnessite. The average 0-0 distance of the tetrahedral layer is 3.00 Å in illite-montmorillonite mixed-layered clays, which differs only 3.4 percent from the 2.90 Å distance of the hexagonally closed-packed oxygens in birnessite..." Varnish formation, then, is a nanometre-scale marriage between clay minerals and the millennial accumulation and decay of Mn-Fe-encrusted bacterial casts. Nanometre-sized Mn-Fe remains of bacteria both feather and cement clay minerals to the rock and to prior varnish.

 Any successful model of varnish formation, however, must also explain two additional issues associated with varnish chemistry.   First, iron is enhanced slightly in varnish over crustal and adjacent soil concentrations. Although most of the iron derives from the clay minerals, some is biotically enhanced (see Fig. 8.7 and Adams et al. 1992; Dorn & Oberlander 1982; Sterflinger et al. 1999) as the nanometer-scale Fe seen in the bacterial casts (Krinsley 1998) is also found reworked into the weathered remnants of clay minerals seen in HRTEM imagery.  Considerable HRTEM work, however, is needed to better document processes of minor iron enhancement.

 Second,  a number of trace elements are enhanced in rock varnish (Engel & Sharp 1958; Lakin et al. 1963; Bard 1979). This enrichment is explained abiotically by Thiagarajan and Lee (2004); trace element enrichment starts when acidic pH values in rainwater ~ 5.7 dissolve rare earths, Co, Ni, Pb, Sr, Rb, Cs and other trace elements from dust particles on rock surfaces.  Co-precipitation of these trace elements with Mn-Fe oxyhydroxides leaves behind the aeolian dust particles that then blow away.  The net process is the sort of concept proposed for the abiotic enhancement of manganese — analogous to a water filter that slowly traps water impurities; in the case of varnish, the dust particles slowly leave behind trace elements trapped by the Mn-Fe minerals.In addition, the opal observed by Perry et al. (2006) could be precipitated as water moves throughout and redistributes varnish constituents (Dorn & Krinsley 1991; Krinsley 1998).

Evaluating Model Strengths and Weaknesses

In addition to the great enhancement of Mn, other criteria have been proposed to adjudicate these different hypotheses (Table 8.4).  No single model explains all observed aspects of Table 8.4, but three of the proposed models run into severe difficulties when they are tasked to explain key varnish characteristics: abiotic enhancement of manganese; lithobionts or their organic remains produce varnish; and silica binding.  These three problematic models, for example, have no severe rate-limiting step.  The processes suggested would all produce rock varnish very rapidly, from 100 to 10,000 times faster than the 1-10 µm/millennia rates observed in nature (Table 8.4).

Table 8.4: Criteria that have been used to adjudicate competing models of rock varnish formation.
Criteria A successful model must be able to explain: ...
Accretion Rate ... typical rates of accretion on the order of a 1-10 µm per millennia (Dorn, 1998; Liu and Broecker, 2000).  Although faster-growing varnishes occur (Dorn and Meek, 1995), varnish accretion rates based on studies of over 104 microbasins (Liu, 2003; Liu, 2006; Liu and Broecker, 2006) demands an extreme rate-limiting step for whatever process makes varnish.
Clay Minerals ... the dominance of clay minerals in rock varnish (Broecker and Liu, 2001; Diaz et al., 2002; Dorn, 1998; Dorn and Oberlander, 1982; Israel et al., 1997; Krinsley, 1998; Krinsley et al., 1990; Krinsley et al., 1995; Potter and Rossman, 1977, 1979c; Probst et al., 2002). 
Fe behavior ... the differential enhancement of iron in different VML (Broecker and Liu, 2001; Liu and Broecker, 2006; Liu et al., 2000) and different places (Adams et al., 1992; Allen et al., 2004; Dorn, 1998).
Laboratory creation ... the creation of artificial varnish coatings (Dorn and Oberlander, 1981, 1982; Jones, 1991; Krumbein and Jens, 1981) may considered by some to be a vital criteria.  However, given the extraordinary time scale jump between any laboratory experiment and natural varnish formation, and the extreme rate-limiting step involved in natural varnish formation, rigid application of this criteria may be problematic.
Lithobionts and organic remains ... the occurrence of different types of lithobionts and the nature of organic remains. The plethora of in situ and cultured organisms, as well as spores, polysaccharides and other humic substances associated with varnish (see review in Dorn, 1998 and since (Allen et al., 2004; Gonzalez et al., 1999; Gorbushina et al., 2002; Kuhlman et al., 2005; Kuhlman et al., 2006; Kurtz and Netoff, 2001; Perry et al., 2002; Perry et al., 2003; Perry and Kolb, 2003; Schelble et al., 2005; Viles and Goudie, 2004) ) must be explained. 
Mn Enhancement ... the enhancement Mn typically more than a factor of fifty above potential source materials (Engel and Sharp, 1958; Lucas, 1905; von Humboldt, 1812). 
Mn-mineralogy ... Mn-mineralogy characteristic of birnessite-family minerals (McKeown and Post, 2001; Potter, 1979; Potter and Rossman, 1979a, b; Probst et al., 2001).  Very difficult to study because of its small size, Mn-oxides in range from large tunnel phases to layer phases such as the birnessite family notable in varnish.
Not just a few samples ... observations at sites around the world. The cost of microanalytical investigations demands analyses of only a few samples at a time. Nonetheless, Oberlander (1994: 118) emphasizes: "researchers should be warned against generalizing too confidently from studies of single localities". 
Paucity of microfossils ... the extremely infrequent occurrence of preserved microfossils. Examination of 104 (Liu, 2003; Liu, 2006; Liu and Broecker, 2006), and decades of research has generated only a few observations of microfossils (Dorn, 1998; Dorn and Meek, 1995; Flood et al., 2003; Krinsley, 1998)
Rock Coating Geography ... why does rock varnish grow in one place and other rock coatings elsewhere. Over a dozen major types of coatings form on terrestrial rock surfaces (Dorn, 1998).  A successful model must be able to why, for example, rock varnish grows with iron films, silica glaze, phosphate skins, and oxalate crusts in the Khumbu of Nepal (Dorn, 1998: 360-361) and with dust films, carbonate crust, phosphate film, silica glaze, and oxalate crusts in Tibet (Dorn, 1998: 367-369).
Rock Varnish Variety and Geography ... why different types of rock varnishes occur where they occur. Rock varnishes classification must be dealt with, a very complex topic summarized in Dorn (1998: 214-224).  Consider just the geography of light. Although light can be a key issue for some rock coatings (McKnight et al., 1988; Nienow et al., 1988), formation of varnish in locales with little or no light (Dorn, 1998; Dorn and Dragovich, 1990; Douglas, 1987; Douglas et al., 1994; Villa et al., 1995; Whalley et al., 1990) places strong constraints on light-related processes.
VML ... the revolution in varnish microlamination (VML) understanding. Over ten thousand 104 sedimentary microbasins analyzed by Liu  (Broecker and Liu, 2001; Liu, 1994; Liu, 2003; Liu, 2006; Liu and Broecker, 2006; Liu and Broecker, 2000; Liu et al., 2000), a method subject to blind testing (Marston, 2003), reveals clear late Pleistocene and Holocene patterns in abundance of major varnish constituents connected to climate change. The characteristics of this single largest varnish data set must be explained.

The viable polygenetic model, in contrast, turns on the very slow processes by which manganese enhancement takes place and very slow processes by which manganese and iron interact with clay minerals.  A significant advantage of the polygenetic model is that it does not suffer from the fatal flaws extant in the other three models (Table 8.5). A minor advantage of the polygenetic model is that it is fully compatible with   processes of the silica binding model and the presence of adventitious lithobionts and their organic remains in varnish.  An issue for the polygenetic model, however, is that the abiotic enhancement hypothesis has not been falsified — a very difficult task given the rate of varnish formation.

Table 8.5: Performance of alternative rock varnish conceptual models with respect to adjudicating criteria detailed in Table 8.4.



Abiotic Mn enhancement  Lithobionts or their remains  Polygenetic clay-bacteria-inorganic interactions Silica binding
Accretion Rate There is no rate limiting step that would slow the accretion rate to fit observations. Even given the lowest dust deposition rates and fewest wetting - acidification events, this process would be 102 faster than observed rates. If only a fraction of the reported organic remains or agencies precipitated Mn and Fe, formation would be more than 103 faster than observed rates. The rare Mn enhancement by a few budding bacteria, observed in situ, would generate observed rates. The silica binding process has no rate limiting step to slow growth. Under this model, there is no reason why decadal growth rates for silica glazes would not also apply to rock varnish.
Clay Minerals Clay minerals are not important, and hence there is no reason why they should dominate in varnish. Clay minerals are not important, and hence there is no reason why they should dominate in varnish. Clay minerals are critical to this model, and the model explains their dominance. Mn mobilized from bacteria fit into octahedral positions, cementing clays. Proponents assert that clay minerals are not important, and hence there is no reason why they should dominate in varnish.
Fe behavior Highly oxidizing conditions in dry climates or in microsites would lead to the observed Fe behavior. Different organisms might favor the observed iron behavior in different conditions and settings. Highly oxidizing conditions in dry climates or in microsites would lead to the observed behavior, since these bacteria do not favor highly oxidizing settings. The observed iron behavior is not important to this process.
Laboratory creation Replication has occurred with similarities and substantial differences. Replication has occurred with similarities and substantial differences. Replication has occurred with similarities and substantial differences. Replication has occurred with similarities and substantial differences.
Lithobionts and organic remains The occurrence of typically less than a percent of organic matter is readily explained by purely adventitious incorporation as Mn enhancement occurs. The occurrence is a prerequisite of this model. The occurrence of less than a percent of organic matter is readily explained by adventitious organics incorporated into varnish. The occurrence of organic matter is predicted by silica binding processes.
Mn Enhancement  This model was developed to explain Mn enhancement. Abundant culturing research documents the presence of Mn-oxidizing organisms. This model explicitly explains Mn enhancement, followed by clay cementation. Mn enhancement is not explained by silica dissolution or reprecipitation.
Mn-mineralogy Mn-mineralogy is not explained. Mn-mineralogy is consistent with biotic enhancement of Mn. This model explains the fit of nanometer-scale Mn-oxides into clay minerals, with detailed HRTEM observations supporting earlier hypotheses (Potter, 1979: 174-175; Potter and Rossman, 1977). Mn-mineralogy is not explained.
Not just a few samples The model's discussion in the literature has been fit to only a few site contexts. Individual studies involve only a very few sites, due to the expenses involved, but cumulative observations have a wide distribution This model has been applied globally and is consistent with Liu's VML observations of over 104 micro-sedimentary basins. The model is too new to have been assessed at sites globally or articulated to the VML data set.
Paucity of microfossils This model would predict the paucity of microfossils. Several studies are notable for presence of microfossils, while others are notable for analysis of organic residues only.  This model explains the paucity of microfossils, because Mn-encrusted cell walls are redistributed at the nanometer scale. Model proponents hope to use microfossils to explore extraterrestrial life, but this makes little sense given poor preservation.
Rock Coating Geography The presence of rock varnish in a plethora of settings that lack repeated oxidation cycles challenges this model.  This model has not yet explained the geography of rock coatings. This model explains the geography of rock coatings (Dorn, 1998). This model has not yet explained why rock varnish or silica glazes exists at one locale and not in other geographic settings.
Rock Varnish Variety and Geography This model has not yet explained the variety and geography of rock varnish. This model has not yet explained the variety and geography of rock varnish. This model explains the variety and geography of rock varnish (chapter 10 in Dorn, 1998). This model requires light and heat. Yet, both are known to be  unnecessary for varnish formation (Anderson and Sollid, 1971; Dorn, 1998; Dorn and Dragovich, 1990; Dorn et al., 1992; Douglas, 1987; Douglas et al., 1994; Villa et al., 1995; Whalley et al., 1990).
VML This model could explain VML observations, whereby conditions of greater alkalinity and fewer wetting events would generate drier VML events and more wetting events would generate wetter VML events. Model proponents have not yet explored the implications of the VML record. This model explains VML observations. Specific HRTEM observations in different VML reveal more nm-scale bacterial fragments in wet periods. Wording to accommodates alternating climatic changes producing VML. However, the silica binding process does not explain VML.


Palaeoenvironmental significance

 Varnish sometimes serves as an indicator of palaeoenvironments even before the Quaternary (Dorn & Dickinson 1989; Marchant et al. 1996).  A buried Miocene hillslope in southern Arizona, for example, hosts a fossilized debris slope where many of the small boulders are covered by fossilized varnish (Dorn & Dickinson 1989).  Simply finding an isolated clast of rock varnish in buried contexts, however, cannot be used to infer any particular past climate (Dorn 1990). But entire slopes or pavements (Marchant et al. 1996) covered with varnished boulders permits palaeoclimatic inherence. In southern Arizona, a reasonable inference is that the occurrence of rock varnish in this buried position (Dorn & Dickinson 1989) indicates an arid to semi-arid climate.

 Microstratigraphic layers in actively forming varnish also provide palaeoenvironmental insights. A wide variety of varnish properties have been used to infer Quaternary paleoenvironmental change.  Micromorphological change (Dorn 1986), trace element geochemistry such as lead profiling (Dorn 1998; Fleisher et al. 1999), oxygen-17 in sulfates (Bao et al. 2001), organic carbon ratios (Dorn et al. 2000), and interlayering with other rock coatings (Dragovich 1986; Dorn 1998) exemplify the potential of varnish microstratigraphy to inform on environmental change.

 One key to extracting palaeoenvironmental data rests in sampling the same type of varnishes at sites used to calibrate the varnish proxy record and unknown sites (Dorn, 1998, p. 214-224).  Another key rests in avoiding post-depositional modification of varnish strata, such as erosion by microcolonial fungi (Fig. 8.3) or cation leaching and redepositional processes (Dorn & Krinsley 1991; Krinsley 1998).

Table 8.6. Different methods that have been used to assess rock varnish chronometry. The refinement of age resolution is either relative, calibrated by independent age control, correlated to discrete events, or numerical based on radiometric measurements.

Method and age resolution Synopsis of method 
Accumulation of Mn and Fe: calibrated age As more varnish accumulates, the mass of manganese and iron gradually increases. Occasionally this old idea is resurrected, but it has long ago been demonstrated to yield inaccurate results in tests against independent control (Bard 1979; Dorn 2000).
Appearance: relative age The appearance of a surface darkens over time as varnish thickens and increases in coverage. However, there are too many exceptions to permit accurate or precise assignment of ages based on visual appearance. For example, varnish can form in under 100 years in selected microenvironments (Dorn & Meek 1995; Krumbein 1969). There is no known method that yields reliable results (Dorn 2000), although individuals have tried to make visual measurements using different techniques. 
Cation-ratio dating: calibrated age Rock varnish contains elements that are leached (washed out) at fast rates and elements that are not leached rapidly (Dorn & Krinsley 1991; Krinsley 1998). Over time, a ratio of leached to immobile declines over time (Dorn 2000).While differences in field and laboratory sampling procedures have yielded problematic results in multiple investigations attempting replication (e.g., Bamforth 1997; Watchman 2000), if the correct type of varnish is used, the method performs well in blind tests (Loendorf 1991). It has been replicated in such places as China (Zhang et al. 1990), Israel (Patyk-Kara et al. 1997), South Africa (Whitley & Annegarn 1994) southern Nevada (Whitney & Harrington 1993), and elsewhere.
Foreign Material Analysis: correlated age Rock carvings made historically may have used steel. The presence of steel remains embedded in a carving would invalidate claims of antiquity, whereas presence of such material as quartz would be consistent with prehistoric age (Whitley et al. 1999). Meticulous sampling and examination of cross-section with electron microprobe is requested to determine if steel remains exist in a petroglyph groove.
Lead Profiles: correlated age 20th century lead pollution is recorded in rock varnish, because the iron and manganese in varnish scavenges lead. This leads to a "spike" in the very surface layer from 20th century pollution. Confidence is high, because the method (Dorn 1998) has been replicated (Fleisher et al. 1999; Thiagarajan & Lee 2004; Hodge et al., 2005) with no publications yet critical of the technique. The method uses sampling and measurement of chemical changes with depth on the scale of microns, analyzed with an instrument such as an electron or ion microprobe.
Organic Carbon Ratio: calibrated age Organic carbon exists in an open system in the rock varnish that covers petroglyphs. This method compares the more mobile carbon and the more stable carbon. The method is best used in soil settings (Harrison & Frink 2000), but it has been applied experimentally to rock varnish in desert pavements (Dorn et al. 2000).The method requires careful mechanical extraction of rock varnish, that is then subjected to basic soil wet chemistry procedures.
Radiocarbon - carbonate: numerical age Calcium carbonate sometimes forms over varnish, and can be radiocarbon dated, providing a minimum age for such features as rock art. The method has been used in Australia (Dragovich 1986) and eastern California (Cerveny et al., 2005).
Radiocarbon - organic: numerical age The hope is that carbon trapped by coating provides minimum age for the petroglyph. First developed in 1986, Watchman (1997) and Dorn (1997) both found the presence of organic carbon that pre-dates and post-dates the exposure of the rock suface ? in this case a blind test of petroglyph dating in Portugal. The only person who still uses organic carbon of unknown residues in radiocarbon dating (Huyge et al. 2001; Watchman 2000), Watchman now admits that he has not tested results against independent controls — despite using the method in a consulting business (Watchman 2002; Whitley & Simon, 2002a,b).
Radiocarbon - oxalate: numerical age The inorganic mineral oxalate (e.g., whewellite: CaC2O4H2O) sometimes deposits on top of or underneath rock varnish. Because this mineral contains datable carbon, the radiocarbon age can provide a minimum age for the underlying or overlying varnish. The most reliable research on radiocarbon dating of oxalates in rock surface contexts has been conducted in west Texas (Rowe 2000; Russ et al. 2000)
Varnish Microlaminations: correlated age Climate fluctuations change the patterns of layers deposited in varnish. The confidence level is high, because the method has been replicated in a rigorous blind test discussed in the text (Marston 2003).

  Of the variety of potential palaeoenvironmental and dating tools yet proposed for varnish (Table 8.6), the most potent method is varnish microlaminations (VML).  Black, yellow, and orange layers in varnish reflect regional climatic changes (Dorn 1990; Cremaschi 1996; Liu & Dorn 1996; Zhou et al. 2000; Broecker and Liu, 2001; Diaz et al., 2002; Lee and Bland, 2003; Liu 2003; Liu 2006), where black layers, rich in manganese, record wet intervals.  Orange and yellow layers with less manganese record drier periods (Jones, 1991). A blind test of the VML method (Liu 2003; Marston 2003), communicated and refereed by the editor of Geomorphology, resulted in this conclusion:

This issue contains two articles that together constitute a blind test of the utility of rock varnish microstratigraphy as an indicator of the age of a Quaternary basalt flow in the Mohave Desert. This test should be of special interest to those who have followed the debate over whether varnish microstratigraphy provides a reliable dating tool, a debate that has reached disturbing levels of acrimony in the literature. Fred Phillips (New Mexico Tech) utilized cosmogenic 36Cl dating, and Liu (Lamont-Doherty Earth Observatory, Columbia University) utilized rock varnish microstratigraphy to obtain the ages of five different flows, two of which had been dated in previous work and three of which had never been dated. The manuscripts were submitted and reviewed with neither author aware of the results of the other. Once the manuscripts were revised and accepted, the results were shared so each author could compare and contrast results obtained by the two methods. In four of the five cases, dates obtained by the two methods were in close agreement. Independent dates obtained by Phillips and Liu on the Cima ‘‘I’’ flow did not agree as well, but this may be attributed to the two authors having sampled at slightly different sites, which may have in fact been from flows of contrasting age. Results of the blind test provide convincing evidence that varnish microstratigraphy is a valid dating tool to estimate surface exposure ages. (Marston 2003) The method works in the western United States (Liu et al. 2000), western China (Zhou et al. 2000), North Africa (Cremaschi 1996), and Patagonia (T. Liu, personal communication, 2001), with the notation that these different regions require different calibrations. The calibrations (Liu 2003; Liu 2006) for the Great Basin of the western United States extends back to before 140,000 years.  Grayscale transformations do not do the method justice, but Fig. 8.10 illustrates that progressively older varnishes display progressively more detailed VML sequences.

Fig. 8.10. Varnish microlaminations (VML) are the most powerful "tool" yet developed in rock varnish research (Liu, 2006).This figure illustrates the potential of this tool to yield unique insights in geomorphic research.
           Upper Part:   Looking at this oblique photograph of the Warm Springs alluvial fan in Death Valley, there would be a tendency for most of today’s desert geomorphologists to believe in the importance of climatic change in fan development (Farr & Chadwick 1996; Harvey et al. 1999). Such thinking, however, is not supported by the VML method. Although VML provides the highest resolution palaeoclimatic information yet available for alluvial fans, this detailed mapping fails to reveal a clear relationship between climatic changes and alluvial-fan development (Dorn 1996). Climatic change may play a role, but this hypothesis is not testable  — even with the highest resolution method of VML. High magnitude and high frequency palaeoclimatic fluctuations occur with greater frequency than can be related to aggradational events. The inset mapping is modified from Liu and Dorn (1996), with the inset satellite image from NASA.
            Lower Part of Figure:   Varnish microlaminations from Death Valley, California, illustrate progressively more complex stratigraphic sequences in late Pleistocene varnishes. The uppermost LU-1 layer appears yellow in ultrathin section. The other light gray layers appear orange.  Dark layers appear black.   This figure was modified from Liu and Dorn (Liu & Dorn 1996) with the calibration based on Liu (2003).

 The great power of the VML method rests in obtaining two types of information at once: palaeoenvironmental sequence and calibrated age.  For example, research on the alluvial fans of Death Valley (Fig. 8.9) permit the first-ever high-resolution assessment of the linkage between fan chronology and palaeoenvironmental fluctuations (Liu & Dorn 1996).  In another example, microlaminations permit explorations of the relationship between archaeology and palaeoclimate (Dorn 2000; Dorn et al. 2000; Dorn, 2006).

 For those interested in applying varnish to Quaternary research, VMLs offer the ability to link geomorphic and archaeologic chronology with regional palaeoenvironmental fluctuations (Liu, 2006).  Before they embark on this methodology, researchers should internalize two issues.  First, the making of ultra-thin sections is extraordinarily difficult; sections a bit too thick are opaque, while sections too thin erode away easily — requiring great patience and a light enough laboratory touch.  Second, Marston’s (2003) careful handling of this blind test of microlaminations, as well as other blind tests (Loendorf 1991; Arrowsmith et al. 1998) underscores the importance of understanding classification (Dorn, 1998, p. 214-224) prior to the sampling phase.

Relationships to other terrestrial geochemical sediments

Rock varnish rarely exists as a discrete geochemical sedimentary deposit.  Aspect (Waragai 1998) and slope position (Palmer 2002), for example, influence metre-scale facies changes.  Vertical interlarying takes place on the scale of microns with every other type of rock coating (Table 8.1).  Examples of interlayering include: with oxalates and lithobiontic coatings in waterflow locales; with nitrate crusts in caves; with silica glaze in desert pavements; with phosphates near locales of decomposition of bird excrement; with carbonate skins in archaeological ground figures; with case hardening agents; with anthropogenic pigment in rock art; or with heavy metal skins in periglacial meltwater streams.  At least 102 publications have thus far examined varnish interlayered with other types of rock coatings — thus unknowingly mixing "apples and oranges" in possible misinterpretations.

 There have been a number of suggestions in the literature (cf. chapters 10 and 13 in Dorn 1998) equating silica glazes, coatings of largely amorphous silica, with abiotic formation of desert varnish.  In silica glaze formation, soluble Al-Si complexes [Al(OSi(OH)3)2+] are first released from the weathering of phyllosilicate minerals. These Al-Si complexes are then stabilized through very gentle wetting such as dew deposition or movement of capillary water on rock surfaces.  Once stabilized, silicic acid and these Al-Si complexes more readily bond the silica glaze coating to the rock (Dorn 1998: 319).  In HRTEM imagery nanometre-scale bands of silica glaze do experience facies changes with varnish and in BSE imagery micrometre-scale bands of silica glaze interlayer with varnish. Thus, there could also be a cementation role for silica glaze in varnish accretion.

 Interlayering of different types of rock coatings, while creating the strong potential for gross misinterpretation, creates excellent opportunities to apply rock coatings to preservation of stone.  Natural settings illustrate the importance of mixing different types of cementing agents (Fig. 8.11). Iron films play the natural analog of artificial bonding agents where sandstone grains are held tightly. Iron films create a harder rock face than a more porous cementing mixture of varnish and silica glaze. However, the intercalation of varnish and silica glaze permits much more interstitial water to exit such lithologies as porous sandstone.  The stronger cementing agent, iron film, enhances decay in the underlying weathering rind (Tratebas et al. 2004).  Whereas spalling under the iron film-hardened sandstone results in rapid erosion of the decayed sandstone, the varnish+silica glaze mixture displays slower and less dramatic erosion (Tratebas et al. 2004).

Fig. 8.11. Rock varnish interdigitates with iron film and silica glaze at Whoopup Canyon, Wyoming.  Iron film (left BSE image) acts as a case hardening agent, and rock varnish accretes on top of the iron film exposed by petroglyph manufacturing.  Varnish actively assists in case hardening (right BSE image) when the leached cations reprecipitate with silica glaze in sandstone pores.

 Yet another example of using the interlayering of varnish with other rock coatings is seen in the prehistoric modification of a landforms. Consider geoglyphs created by the disturbance of a desert pavement.  Boulders are often piled, reorienting the natural varnish sequence.  Finding subaerial black varnish shoved into the soil results in the "freezing" of the varnish microlaminations sequence (Liu 2003) underneath the superposition of a soil-contact iron film. At the same time, subaerial microlaminations form on top of the orange varnish.  Boulders thrust far enough into the ground may form calcrete, material with potential for understanding the palaeoenvironment (Deutz et al. 2001) and timing (Chen & Polach 1986) of the disturbance.  Thus, understanding all variations of rock varnish and its interdigitation with other rock coatings creates immense possibilities in analyzing ancient modifications to stone (N. Cerveny, personal communication, 2004).

Summary and directions for future research

Desert varnish is a rock coating that is brown to black colour and is characterized by clay minerals cemented to rock surfaces by oxyhydroxides of manganese and iron. The individual Mn-Fe minerals are nanometre in size, a result of the decay of bacterial casts.  These remnants of bacteria dissolve and reprecipitate in clay minerals — cementing the mix of clays and oxyhydroxides to rock surfaces. Since varnish forms in all terrestrial environments, "rock varnish" is often used as a synonym, but the term "desert varnish" is common because varnishes are most geochemically stable in arid regions. The constituents in varnish accrete on the host rock with thicknesses typically less than 0.1 millimetres.  Usually dull in lustre, its occasional sheen comes from a smooth surface micromorphology in combination with manganese enrichment at the very surface of the varnish. The most useful application of rock varnish rests in the layering pattern of black (manganese-rich) layers and orange (manganese-poor) layers, that reveal past environmental changes experienced by rock surfaces.

 Rock varnish is beginning to see an infusion of dollars related to the search for life on Mars by NASA (DiGregorio 2002), using the assumption that biogenically-formed rock coatings on Earth represents a possible Martian analog (Israel et al. 1997; Johnson et al. 2001; Probst et al. 2001; Gorbushina et al. 2002; Wierzchos et al. 2003; Allen et al. 2004; Kuhlman et al., 2005, 2006; Perry et al. 2006). One need in this research arena rests in understanding the transition between Mn-Fe-casts of Mn-oxidizing organisms on varnish surfaces to their preservation within varnish, and onto their remobilization into a clay matrix.  Cold-dry contexts such as the Tibetan Plateau (Dorn 1998) and Antarctica (Dorn et al. 1992; Krinsley 1998) reveals evidence of this process. Simply studying the remains of organisms, most of which are purely adventitious, in warm deserts will inevitably lead to misinterpretion and research dead ends.  Exploring how biogenic remains in the coldest and driest terrestrial settings translate into biogenic coatings, again, requires a penchant towards HRTEM studies. The real key in astrobiology research as it relates to varnish, however, does not rest in understanding terrestrial varnish; the biggest gap is the need to falsify abiotic explanations of Mn enhancement on Mars.  Imagine when the astrobiology community discovers Mn-rich rock coatings on Mars, claiming evidence of life — only to encounter the reality that simple mechanisms of abiotic enhancement of Mn have not been falsified for Mars.

The greatest need for future research in terrestrial varnish research rests with individuals who are able to carry out extraordinary care in the preparation of laboratory samples such as microlaminations (Liu, 2003; Marston, 2003). This powerful varnish technique involves Pleistocene (Liu, 2003) and Holocene microlamination sequences (Liu and Broecker, 2006) that exist in rapidly forming varnishes and can be found in microenvironments of greater moisture in a largely arid setting (Dorn et al., 2000; Broecker and Liu, 2001).  Robust calibrations of these sequences now provide archaeologists and geomorphologists with a powerful tool to understand the interface of chronology and palaeoenvironmental changes over the last few millennia (Liu, 2006).

 Another need rests in linking varnishes seen at the satellite remote sensing scale (White 1993; Farr & Chadwick 1996; Milana 2000; Kenea 2001) to centimeter-scale remote sensing sensors placed in such platforms as balloons or helicopters.  Bridging scales between ground and satellite (Fig. 8.4) remains a giant void.  Palmer (2002) illustrates the potential in quantitative landform-scale catenas of spatial variability of rock coatings.  Certainly, such field work must be calibrated by textural and chemical analyses using such equipment as an electron microprobe.  However, a core research question that remains unresolved is a detailed exploration of spatial heterogeneity of varnish on the scale of single landforms — a topic brought in focus by the realization that 99%  of varnish publications are based on only a few "grab bag" samples.  Field-based (with laboratory calibration) studies of varnish spatial heterogeneity will be needed to create the solid the backbone of future advances in varnish research.

Acknowledgements. Special thanks to the late James Clark for assistance in electron microscopy. Research was supported by in part by sabbatical support from ASU.

Adams, J.B., Palmer, F. & Staley, J.T. (1992) Rock weathering in deserts: mobilization and concentration of ferric iron by microorganisms. Geomicrobiology Journal 10, 99-114.

Allen, C., Probst, L.W., Flood, B.E., Longazo, T.G., Scheble, R.T. & Westall, F. (2004) Meridiani Planum hematite deposit and the search for evidence of life on Mars—iron mineralization of microorganisms in rock varnish. Icarus  171, 20-30.

Anderson, J.L. & Sollid, J.L. (1971) Glacial chronology and glacial geomorphology in the marginal zones of the glaciers, Midtadlsbreen and Nigardsbreen, South Norway. Norsk Geografisk Tidsskrift 25, 1-35.

Arrowsmith, J. R., Rice, G. E. & Hower, J. C. (1998) Documentation of carbon-rich fragments in varnish-covered rocks from central Arizona using electron microscopy and coal petrography. Submitted to Science.

Bamforth, D. (1997) Cation-ratio dating and archaeological research design: A response to Harry. American Antiquity 62, 121-129

Bao, H., Michalski, G. M. & Thiemens, M. H. (2001) Sulfate oxygen-17 anomalies in desert varnishes. Geochimica et Cosmochimica Acta 65, 2029-2036.

Bard, J. C. (1979) The development of a patination dating technique for Great Basin petroglyphs utilizing neutron activation and X-ray fluorescence analyses. Ph.D. Dissertation. Department of AnthropologyUniversity of California, Berkeley.

Borns, D. J., Adams, J. B., Curtiss, B., Farr, B., Palmer, T., Staley, J. & Taylor-George, S. (1980) The role of microorganisms in the formation of desert varnish and other rock coatings: SEM study. Geological Society of America Abstracts with Program 12 (7), 390.

Bull, W. B. (1984) Alluvial fans and pediments of southern Arizona.  In Smiley, T. L., Nations, J. D., Péwé, T. L. & Schafer, J. P. (eds.) Landscapes of Arizona. The geological story, University Press of America, New York, pp. 229-252.

Broecker, W. S. & Liu, T. (2001) Rock varnish: recorder of desert wetness? GSA Today 11 (8), 4-10.

Cahill, T. A. (1986) Particle-Induced X-ray Emission.  In Whan, R. E. (eds.) Metals Handbook Ninth Edition. Volume 10. Materials Characterization, American Society for Metals, Metals Park, pp. 102-108.

Cerveny, N. Kaldenberg, R., Reed, J. Whitley, D.S., Simon, J. & Dorn, R.I. (2006). A new strategy for analyzing the chronometry of constructed rock features in deserts. Geoarchaeology 21: 181-203.

Chafetz, H. S., Akdim, B., Julia, R. & Reid, A. (1998) Mn- and Fe-rich black travertine shrubs: Bacterially (and nanobacterially) induced precipitates. Journal of Sedimentary Research 68, 404-412.

Chen, Y. & Polach, J. (1986) Validity of C-14 ages of carbonates in sediments. Radiocarbon 28 (2A), 464-472.

Coudé-Gaussen, G., Rognon, P. & Federoff, N. (1984) Piegeage de poussières éoliennes dans des fissures de granitoides due Sinai oriental. Compte Rendus de l'Academie des Sciences de Paris II (298), 369-374.

Cremaschi, M. (1996) The desert varnish in the Messak Sattafet (Fezzan, Libryan Sahara), age, archaeological context and paleo-environmental implication. Geoarchaeology 11, 393-421.

Curtiss, B., Adams, J.B. & Ghiorso, M.S. (1985) Origin, development and chemistry of silica-alumina rock coatings from the semiarid regions of the island of Hawaii. Geochemica et Cosmochimica Acta 49, 49-56.

Deutz, P., Montanez, I.P., Monger, H.C. & Morrison, J. (2001) Morphology and isotope heterogeneity of Late Quaternary pedogenic carbonates:  Implications for paleosol carbonates as paleoenvironmental proxies. Palaeogeography Palaeoclimatology Palaeoecology 166, 293-317.

Diaz, T.A., Bailley, T.L. & Orndorff, R.L. (2002) SEM analysis of vertical and lateral variations in desert varnish chemistry from the Lahontan Mountains, Nevada. Geological Society of America Abstracts with Programs, May 7-9 Meeting, <>.

DiGregorio, B. E. (2002) Rock varnish as a habitat for extant life on Mars.  In Hoover, R. B., Levin, G. V., Paepe, R. R. & Rozanov, A. Y. (eds.) Instruments, Methods, and Missions for Astrobiology IV, NASA, SPIE Vol. 4495, pp. 120-130.

Dorn, R. I. (1986) Rock varnish as an indicator of aeolian environmental change.  In Nickling, W. G. (eds.) Aeolian Geomorphology, Allen & Unwin, London, pp. 291-307.

Dorn, R. I. (1990) Quaternary alkalinity fluctuations recorded in rock varnish microlaminations on western U.S.A. volcanics. Palaeogeography, Palaeoclimatology, Palaeoecology 76, 291-310.

Dorn, R. I. (1994) Dating petroglyphs with a 3-tier rock varnish approach.  In Whitley, D. S. & Loendorf, L. (eds.) New Light on Old Art: Advances in Hunterer-Gatherer Rock Art Research., UCLA Inst. Arch. Monograph Ser.36, pp. 12-36.

Dorn, R. I. (1996) Climatic hypotheses of alluvial-fan evolution in Death Valley are not testable.  In Rhoads, B. L. & Thorn, C. E. (eds.) The Scientific Nature of Geomorphology, Wiley, New York, pp. 191-220.

Dorn, R.I. (1997) Constraining the age of the Côa valley (Portugal) engravings with 14C. Antiquity 71, 105-115.

Dorn, R. I. (1998) Rock coatings, Elsevier, Amsterdam.

Dorn, R. I. (2000) Chronometric techniques: Engravings.  In Whitley, D. S. (eds.) Handbook of rock art research, Altamira Press, Walnut Creek, pp. 167-189.

Dorn, R.I. (2006) Petroglyphs in Petrified Forest National Park: Role of rock coatings as agents of sustainability and as indicators of antiquity. Bulletin of Museum of Northern Arizona 63, 52-63.

Dorn, R. I., Cahill, T. A., Eldred, R. A., Gill, T. E., Kusko, B., Bach, A. & Elliott-Fisk, D. L. (1990) Dating rock varnishes by the cation ratio method with PIXE, ICP, and the electron microprobe. International Journal of PIXE 1, 157-195.

Dorn, R. I. & Dickinson, W. R. (1989) First paleoenvironmental interpretation of a pre-Quaternary rock varnish site, Davidson Canyon, south Arizona. Geology 17, 1029-1031.

Dorn, R.I. & Dragovich, D. (1990) Interpretation of rock varnish in Australia: Case studies from the Arid Zone. Australian Geographer 21, 18-32.

Dorn, R. I. & Krinsley, D. H. (1991) Cation-leaching sites in rock varnish. Geology 19, 1077-1080.

Dorn, R. I., Krinsley, D. H., Liu, T., Anderson, S., Clark, J., Cahill, T. A. & Gill, T. E. (1992) Manganese-rich rock varnish does occur in Antarctica. Chemical Geology 99, 289-298.

Dorn, R. I. & Meek, N. (1995) Rapid formation of rock varnish and other rock coatings on slag deposits near Fontana. Earth Surface Processes and Landforms 20, 547-560.

Dorn, R. I. & Oberlander, T. M. (1981) Microbial origin of desert varnish. Science 213, 1245-1247.

Dorn, R. I. & Oberlander, T. M. (1982) Rock varnish. Progress in Physical Geography 6, 317-367.

Dorn, R. I., Stasack, E., Stasack, E. & Clarkson, P. (2000) Through the looking glass: Analyzing petroglyphs and geoglyphs with different perspectives. American Indian Rock Art 27, 77-96

Dragovich, D. (1986) Minimum age of some desert varnish near Broken Hill, New South Wales. Search 17, 149-151.

Dragovich, D. (1993) Distribution and chemical composition of microcolonial fungi and rock coatings from arid Australia. Physical Geography 14, 323-341.

Douglas, G.R. (1987) Manganese-rich rock coatings from Iceland. Earth Surface Processes and Landforms 12, 301-310.

Douglas, G.R., McGreevy, J.P. & Whalley, W.B. (1994) Mineralogical aspects of crack development and freeface activity in some basalt cliffs, County Antrim, Northern Ireland.  In Robinson, D.A. & Williams, R.B.G. (eds.) Rock weathering and landform evolution, Wiley, Chichester, pp. 71-88.

Engel, C. G. & Sharp, R. S. (1958) Chemical data on desert varnish. Geological Society of America Bulletin 69, 487-518.

Farr, T. G. & Chadwick, O. A. (1996) Geomorphic processes and remote sensing signatures of alluvial fans in the Kun Lun mountains, China. Journal of Geophysical Research-Planets 101 (E10), 23091-23100.

Ferris, F. G., Beveridge, T. J. & Fyfe, W. S. (1986) Iron-silica crystallite nucleation by bacteria in a geothermal sediment. Nature 320, 609-611.

Fleisher, M., Liu, T., Broecker, W. & Moore, W. (1999) A clue regarding the origin of rock varnish. Geophysical Research Letters 26 (1), 103-106.

Flood, B.E., Allen, C. & Longazo, T. (2003) Microbial fossils detected in desert varnish. Astrobiology 2 (4) 608.

Forbes, E. A., Posner, A. M. & Quirk, J. P. (1976) The specific adsorption of divalent Cd, Co, Cu, Pb, and Zn on goethite. Journal of Soil Science 27, 154-166.

Francis, W. D. (1921) The origin of black coatings of iron and manganese oxides on rocks. Proceedings of the Royal Society of Queensland 32, 110-116.

Glasby, G. P., McPherson, J. G., Kohn, B. P., Johnston, J. H., Keys, J. R., Freeman, A. G. & Tricker, M. J. (1981) Desert varnish in southern Victoria Land, Antarctica. New Zealand Journal of Geology and Geophysics 24, 389-397.

Gonzalez, I., Laiz, L., Hermosin, B., Caballero, B., Incerti, C. & Saiz-Jimenez, C. (1999) Bacteria isolated from rock art paintings: the case of Atlanterra shelter (south Spain). Journal of Microbiological Methods 36 (1-2), 123-127.

Glazovskaya, M.A. (1958) Weathering and primary soil formations in Antarctica. Scientific Paper of the Institute, Moscow University, Faculty of Geography 1, 63-76.

Gorbushina, A.A., Krumbein, W.E. & Volkmann, M. (2002) Rock surfaces as life indicators: New ways to demonstrate life and traces of former life. Astrobiology 2 (2), 203-213.

Gordon, S.J. & Dorn, R.I. (2005) Localized weathering: Implications for theoretical and applied studies. Professional Geographer 57, 28-43.

Haberland, W. (1975) Untersuchungen an Krusten, Wustenlacken und Polituren auf Gesteinsoberflachen der nordlichen und mittlerent Saharan (Libyen und Tchad). Berliner Geographische Abhandlungen 21, 1-77.

Harrison, R. & Frink, D.S. (2000) The OCR carbon dating procedure in Australia: New dates from Wilinyjibari Rockshelter, southeaster Kimberley, Western Australia. Australian Archaeology 51, 6-15.

Harvey, A. M., Wigand, P. E. & Wells, S. G. (1999) Response of alluvial fan systems to the late Pleistocene to Holocene climatic transition: contrasts between the margins of pluvial Lakes Lahontan and Mojave, Nevada and California, USA. Catena 36, 255-281.

Hem, J. D. & Lind, C. J. (1983) Nonequilibrium models for predicting forms of precipitated manganese oxides. Geochimica Cosmochimica Acta 47, 2037-2046.

Hobbs, W. H. (1918) The peculiar weathering processes of desert regions with illustrations from Egypt and the Soudan. Michigan Academy of Sciences Annual Report 20, 93-98.

Hodge, V.F., Farmer, D.E., Diaz, T. & Orndorff, R.L. (2005) Prompt detection of alpha particles from 210Po: another clue to the origin of rock varnish? Journal of Environmental Radioactivity 78, 331-342.

Hooke, R. L., Yang, H. & Weiblen, P. W. (1969) Desert varnish: an electron probe study. Journal Geology 77, 275-288.

Howard, C. D. (2002) The gloss patination of flint artifacts. Plains Anthropologist 47, 283-287.

Hungate, B., Danin, A., Pellerin, N. B., Stemmler, J., Kjellander, P., Adams, J. B. & Staley, J. T. (1987) Characterization of manganese-oxidizing (MnII-->MnIV) bacteria from Negev Desert rock varnish: implications in desert varnish formation. Canadian Journal Microbiology 33, 939-943.

Hunt, A.G. & J.Q. Wu. (2004) Climatic influences on Holocene variations in soil erosion rates on a small hill in the Mojave Desert. Geomorphology 58, 263-289.

Huyge, D., Watchman, A., De Dapper, M. & Marchi, E. (2001) Dating Egypt's oldest 'art': AMS 14C age determinations of rock varnishes covering petroglyphs at El-Hosh (Upper Egypt). Antiquity 75,68-72.

Ishimaru, S. & Yoshikawa, K. (2000) The weathering of granodiorite porphyry in the Thiel Mountains, Inland Antarctica. Geografiska Annaler A 82, 45-57.

Israel, E. J., Arvidson, R. E., Wang, A., Pasteris, J. D. & Jolliff, B. L. (1997) Laser Raman spectroscopy of varnished basalt and implications for in situ measurements of Martian rocks. Journal of Geophysical Research-Planets 102 (E12), 28705-28716.

Jenne, E. A. (1968) Controls on Mn, Fe, Co, Ni, Cu and Zn concentrations in soils and water: the significant role of hydrous Mn and Fe oxides.  In Gould, R. F. (eds.) Trace Inorganics in Water, American Chemical Society, Washington D.C., pp. 337-387.

Jenny, H. (1941) Factors of Soil Formation, McGraw-Hill, New York.

Johnson, J. R., Ruff, S. W., Moersch, J., Roush, T., Horton, K., Bishop, J., Cabrol, N. A., Cockell, C., Gazis, P., Newsom, H. E. & Stoker, C. (2001) Geological characterization of remote field sites using visible and infrared spectroscopy: Results from the 1999 Marsokhod field test. Journal of Geophysical Research - Planets 106 (E4), 7683-7711.

Jones, C.E. (1991) Characteristics and origin of rock varnish from the hyperarid coastal deserts of northern Peru. Quaternary Research 35, 116-129.

Junta, J.L. & Hochella Jr. M.F. (1994) Manganese (II) oxidation at mineral surfaces: A microscopic and spectroscopic study. Geochimica Cosmochimica Acta 48, 897-902.

Kenea, N. H. (2001) Influence of desert varnish on the reflectance of gossans in the context of Landsat TM data, southern Red Sea Hills, Sudan. International Journal of Remote Sensing 22, 1879-1894.

Kennedy, C. B., Scott, S. D. & Ferris, F. G. (2003) Characterization of bacteriogenic iron oxide deposits from Axial Volcano, Juan de Fuca Ridge, Northeast Pacific Ocean. Geomicrobiology Journal 20, 199-214.

Khak-mun, T. (1966) Iron- and manganese-oxidizing microorganisms in soils of South Sakhalin. Microbiology 36 (2), 276-281.

Konhauser, K. O., Fyfe, W. S., Schultze-Lam, S., Ferris, F. G. & Beveridge, T. J. (1994) Iron phosphate precipitation by epilithic microbial biofilms in Arctic Canada. Canadian Journal Earth Science 31, 1320-1324.

Krauskopf, K. B. (1957) Separation of manganese from iron in sedimentary processes. Geochimica et Cosmochimica Acta 12, 61-84.

Krinsley, D. (1998) Models of rock varnish formation constrained by high resolution transmission electron microscopy. Sedimentology 45, 711-725.

Krinsley, D., Dorn, R. I. & Anderson, S. (1990) Factors that may interfere with the dating of rock varnish. Physical Geography 11, 97-119.

Krinsley, D. H., Dorn, R. I. & Tovey, N. K. (1995) Nanometer-scale layering in rock varnish: implications for genesis and paleoenvironmental interpretation. Journal of Geology 103, 106-113.

Krumbein, W. E. (1969) Über den Einfluss der Mikroflora auf die Exogene Dynamik (Verwitterung und Krustenbildung). Geologische Rundschau 58, 333-363.

Krumbein, W. E. (1971) Biologische Entstehung von wüstenlack. Umschau 71, 210-211.

Krumbein, W. E. & Jens, K. (1981) Biogenic rock varnishes of the Negev Desert (Israel): An ecological study of iron and manganese transformation by cyanobacteria and fungi. Oecologia 50, 25-38.

Kuhlman, K.R., Allenbach, L.B., Ball, C.L., Fusco, W.G., La Duc, M.T., Kuhlman, G.M., Anderson, R.C., Stuecker, T., Erickson, I.K., Benardini, J. & Crawford, R.L. (2005). Enumeration, isolation, and characterization of ultraviolet (UV-C) resistant bacteria from rock varnish in the Whipple Mountains, California, Icarus 174, 585-595.

Kuhlman, K.R., Fusco, W.G., Duc, M.T.L., Allenbach, L.B., Ball, C.L., Kuhlman, G. M., Anderson, R.C., Erickson, K., Stuecker, T., Benardini, J., Strap, J.L. & Crawford, R.L. (2006) Diversity of microorganisms within rock varnish in the Whipple Mountains, California. Applied and Environmental Microbiology 72, 1708-1715.

Kurtz, H.D. & Netoff, D.I. (2001) Stabilization of friable sandstone surfaces in a desiccating, wind-abraded environment of south-central Utah by rock surface microorganisms. Journal of Arid Environments 48 (1), 89-100.

Lakin, H.W., Hunt, C.B., Davidson, D.F. & Odea, U. (1963) Variation in minor-element content of desert varnish. U.S. Geological Survey Professional Paper 475-B, B28-B31.

Laudermilk, J. D. (1931) On the origin of desert varnish. American Journal of Science 21, 51-66.

Lee, M.R. & Bland, P.A. (2003) Dating climatic change in hot deserts using desert varnish on meteorite finds. Earth and Planetary Science Letters 206, 187-198.

Linck, G. (1901) Über die dunkelen Rinden der Gesteine der Wüste. Jenaische Zeitschrift für Naturwissenschaft 35, 329-336.

Linck, G. (1928) Über Schutzrinden. Chemie die Erde 4, 67-79.

Liu, T. (1994) Visual microlaminations in rock varnish: a new paleoenvironmental and geomorphic tool in drylands. Ph.D. Dissertation. Department of GeographyArizona State University, Tempe

Liu, T. (2003) Blind testing of rock varnish microstratigraphy as a chronometric indicator: results on late Quaternary lava flows in the Mojave Desert, California. Geomorphology 53, 209-234.

Liu, T. (2006). VML Dating Lab, <accessed June 24, 2006>

Liu, T. & Broecker, W. S. (2000) How fast does rock varnish grow? Geology 28, 183-186.

Liu, T. & Broecker, W. (2006) Holocene rock varnish microstratigraphy and its chronometric application in drylands of western USA. Geomorphology in press.

Liu, T., Broecker, W. S., Bell, J. W. & Mandeville, C. (2000) Terminal Pleistocene wet event recorded in rock varnish from the Las Vegas Valley, southern Nevada. Palaeogeography, Palaeoclimatology, Palaeoecology 161, 423-433.

Liu, T. & Dorn, R. I. (1996) Understanding spatial variability in environmental changes in drylands with rock varnish microlaminations. Annals of the Association of American Geographers 86, 187-212.

Loendorf, L. L. (1991) Cation-ratio varnish dating and petroglyph chronology in southeastern Colorado. Antiquity 65, 246-255.

Lucas, A. (1905) The blackened rocks of the Nile cataracts and of the Egyptian deserts, National Printing Department, Cairo.

Mabbutt, J. A. (1979) Pavements and patterned ground in the Australian stony deserts. Stuttgarter Geographische Studien 93, 107-123.

Mancinelli, R.L., Bishop, J.L. & De, S. (2002) Magnetite in desert varnish and applications to rock varnish on Mars. Lunar and Planetary Science  33, 1046.pdf.

Marchant, D. R., Denton, G. H., Swisher, C. C. & Potter, N. (1996) Late Cenozoic Antarctic paleoclimate reconstructed from volcanic ashes in the Dry Valleys region of southern Victoria Land. Geological Society of America Bulletin 108, 181-194.

Marston, R. A. (2003) Editorial note. Geomorphology 53, 197.

McFadden, L. D., Ritter, J. B. & Wells, S. G. (1989) Use of multiparameter relative-age methods for age estimation and correlation of alluvial fan surfaces on a desert piedmont, eastern Mojave Desert. Quaternary Research 32, 276-290.

McKeown, D. A. & Post, J. E. (2001) Characterization of manganese oxide mineralogy in rock varnish and dendrites using X-ray absorption spectroscopy. American Mineralogist 86, 701-713.

McKnight, D.M., Kimball, B.A. & Bencala, K.E. (1988) Iron photoreduction and oxidation in an acidic mountain stream. Science 240, 637-640.

Milana, J.P. (2000) Characterization of alluvial bajada facies distribution using TM imagery. Sedimentology 47, 741-760.

Morgan, J.J. & Stumm, W. (1965) The role of mulvaleng metal oxides in limnological transofrmations as exemplified by iron and manganese. In Jaag, O. (ed.) Second Water Pollution Research Conference 1, Pergamon Press, New York, pp. 103-131.

Moore, C.B. & Elvidge, C.D. (1982) Desert varnish.  In Bender, G. L. (ed.) Reference handbook on the deserts of North America, Greenwood Press, Westport, pp. 527-536.

Nagy, B., Nagy, L.A., Rigali, M.J., Jones, W.D., Krinsley, D.H. & Sinclair, N. (1991) Rock varnish in the Sonoran Desert: microbiologically mediated accumulation of manganiferous sediments. Sedimentology 38, 1153-1171.

Nienow, J.A., McKay, C.P. & Friedmann, E.I. (1988) The cryptoendolithic microbial environment in the Ross Desert of Antarctica: light in the photosynthetically active region. Microbial Ecology 16, 271-289.

Nishiizumi, K., Kohl, C., Arnold, J., Dorn, R., Klein, J., Fink, D., Middleton, R. & Lal, D. (1993) Role of in situ cosmogenic nuclides 10Be and 26Al in the study of diverse geomorphic processes. Earth Surface Processes and Landforms 18, 407-425.

Oberlander, T.M. (1994) Rock varnish in deserts. In Abrahams, A. & Parsons, A. (eds.) Geomorphology of Desert Environments, Chapman and Hall, London, pp. 106-119.

O'Hara, P., Krinsley, D.H. & Anderson, S.W. (1989) Elemental analysis of rock varnish using the ion microprobe. Geological Society America Abstracts with Programs 21, A165.

Palmer, E. (2002) Feasibility and implications of a rock coating catena: analysis of a desert hillslope. M.A. Thesis. Department of GeographyArizona State University, Tempe.

Palmer, F.E., Staley, J.T., Murray, R.G.E., Counsell, T. & Adams, J.B. (1985) Identification of manganese-oxidizing bacteria from desert varnish. Geomicrobiology Journal 4, 343-360.

Patyk-Kara, N.G., Gorelikova, N.V., Plakht, J., Nechelyustov, G.N. & Chizhova, I.A. (1997) Desert varnish as an indicator of the age of Quaternary formations (Makhtesh Ramon Depression, Central Negev). Transactions (Doklady) of the Russian Academy of Sciences/Earth Science Sections 353A, 348-351.

Peck, S.B. (1986) Bacterial deposition of iron and manganese oxides in North American caves. NSS Bulletin 48, 26-30.

Perel'man, A.I. (1961) Geochemical principles of landscape classification. Soviet Geography Review and Translation 11 (3), 63-73.

Perfil'ev, B.V., Gabe, D.R., Gal'perina, A.M., Rabinovich, V.A., Sapotnitskii, A.A., Sherman, É.É. & Troshanov, É.P. (1965) Applied capillary microscopy. The role of microorganisms in the formation of iron-manganese deposits, Consultants Bureau, New York.

Perry, R.S., Dodsworth, J., Staley, J.T. & Gillespie, A. (2002) Molecular analyses of microbial communities in rock coatings and soils from Death Valley, California. Astrobiology  2(4): 539.

Perry, R.S., Engel, M., Botta, O. & Staley, J.T. (2003) Amino acid analyses of desert varnish from the Sonoran and Mojave deserts. Geomicrobiology Journal 20, 427-438.

Perry, R.S. & Kolb, V.M. (2003) Biological and organic constituents of desert varnish: Review and new hypotheses. In Hoover, R.B. & Rozanov, A.Y. (eds.) Instruments, methods, and missions for Astrobiology VII, SPIE, Bellingham, pp. 202-217.

Perry, R.S., Lynne, B.Y., Sephton, M.A., Kolb, V.M., Perry, C.C. & Staley, J.T. (2006) Backing black opal in the desert sun: The importance of silica in desert varnish. Geology 34, 737-540.

Peterson, F. (1981) Landforms of the Basin and Range Province, defined for soil survey. Nevada Agricultural Experiment Station Technical Bulletin 28.

Phillips, F. M., Zreda, M. G., Plummer, M. A., Benson, L. V., Elmore, D. & Sharma, P. (1996) Chronology for fluctuations in Late Pleistocene Sierra Nevada glaciers and lakes. Science  274, 749-751.

Potter, R. M. (1979) The tetravalent manganese oxides: clarification of their structural variations and relationships and characterization of their occurrence in the terrestrial weathering environment as desert varnish and other manganese oxides. Ph.D. Dissertation. Earth and Planetary ScienceCalifornia Institute of Technology, Pasadena.

Potter, R. M. & Rossman, G. R. (1977) Desert varnish: The importance of clay minerals. Science 196, 1446-1448.

Potter, R. M. & Rossman, G. R. (1978) Manganese oxides in the terrestrial weathering environment. Geological Society of America Abstracts with Programs 10, 473.

Potter, R. M. & Rossman, G. R. (1979a) The manganese- and iron-oxide mineralogy of desert varnish. Chemical Geology 25, 79-94.

Potter, R. M. & Rossman, G. R. (1979b) Mineralogy of manganese dendrites and coatings. American Mineralogist 64, 1219-1226.

Potter, R. M. & Rossman, G. R. (1979c) The tetravalent manganese oxides: identification, hydration, and structural relationships by infrared spectroscopy. American Mineralogist  64, 1199-1218.

Probst, L., Thomas-Keprta, K. & Allen, C. (2001) Desert varnish: preservation of microfossils and biofabric (Earth and Mars?). GSA Abstracts With Programs

Probst, L.W., Allen, C.C., Thomas-Keprta, K.L., Clemett, S.J., Longazo, T.G., Nelman-Gonzalez, M.A. & Sams, C. (2002) Desert varnish - preservation of biofabrics and implications for Mars. Lunar and Planetary Science 33, 1764.pdf.

Reneau, S.L.(1993) Manganese accumulation in rock varnish on a desert piedmont, Mojave Desert, Califonria, and application to evaluating varnish development. Quaternary Research 40, 309-317.

Robert, M., Hardy, M., Elsass, F. & Righi, D. (1990) Genesis, crytallochemistry and organization of soil clays derived from different parent materials in temperate regions. International Conference of Soil Science, Fourteenth 7, 42-47.

Robert, M. & Tessier, D. (1992) Incipient weathering: some new concepts on weathering, clay formation and organization.  In Martini, I. P. & Chesworth, W. (eds.) Weathering, Soils & Paleosols, Amsterdam, Elsevier, pp. 71-105.

Rowe, M.W. (2000) Dating by AMS radiocarbon analysis.  In Whitley, D. S. (eds.) Handbook of rock art research, Altamira Press, Walnut Creek, pp. 139-166.

Russ, J., Loyd, D.H., Boutton, T.W. (2000) A paleoclimate reconstruction for southwestern Texas using oxalate residue from lichen as a paleoclimate proxy. Quaternary International 67,29-36.

Scheffer, F., Meyer, B. & Kalk, E. (1963) Biologische ursachen der wüstenlackbildung. Zeitschrift für Geomorphologie 7, 112-119.

Schelble, R., McDonald, G., Hall, J. & K. Nealson. (2005). Community structure comparison using FAME analysis of desert varnish and soil, Mojave Desert, California. Geomicrobiology Journal 22, 353-360.

Smith, B. J. & Whalley, W. B. (1988) A note on the characteristics and possible origins of desert varnishes from southeast Morocco. Earth Surface Processes and Landforms 13, 251-258.

Staley, J. T., Adams, J. B., Palmer, F., Long, A., Donahue, D. J. & Jull, A. J. T. (1991) Young 14Carbon ages of rock varnish coatings from the Sonoran Desert. Unpublished Manuscript.

Staley, J. T., Jackson, M. J., Palmer, F. E., Adams, J. C., Borns, D.J., Curtiss, B. & Taylor-George, S. (1983) Desert varnish coatings and microcolonial fungi on rocks of the Gibson and Great Victoria Desert, Australia. BMR Journal of Australian Geology and Geophysics 8, 83-87.

Staley, J. T., Palmer, F. & Adams, J. B. (1982) Microcolonial fungi: common inhabitants on desert rocks? Science 215, 1093-1095.

Sterflinger, K., Krumbein, W.E., Lallau, T., & Rullkötter, J. (1999) Microbially mediated orange patination of rock surfaces. Ancient Biomolecules 3, 51-65.

Tebo, B.M., Bargar, J.R., Clement, B.G., Dick, G.J. Murray, K.J. Parker, D., Verity, R. & Webb, S.M. Biogenic manganese oxides: Properties and mechanisms of formation. Annual Review Earth and Planetary Science 32, 287-328.

Tebo, B.M., Ghiorse, W.C., van Waasbergen, L.G., Siering, P.L. & Caspi, R. (1997) Bacterially mediated mineral formation: Insights into manganese (II) oxidation from molecular genetic and biochemical studies. Reviews in Mineralogy 35, 225-266.

Tebo, B.M. & He, L.M. (1998) Microbially mediated oxidative precipitation reactions. In Sparks, D.L. & Grundl, T.J. (eds.) Mineral-Water Interfacial Reactions, American Chemical Society, Washington D.C., pp. 393-414.

Tebo, B.M.,, Johnson, H.A., McCarthy, J.K. & Templeton. A.S. (2005). Geomicrobiology of manganese (II) oxidation. TRENDS in Microbiology 13,  421-428.

Thiagarajan, N. & Lee, C. A. (2004) Trace-element evidence for the origin of desert varnish by direct aqueous atmospheric deposition Earth and Planetary Science Letters 224, 131-141.

Tratebas, A., Cerveny, N. & Dorn, R.I. 2004. The effects of fire on rock art: Microscopic evidence reveals the importance of weathering rinds.  Physical Geography 25, 313-333.

Viles, H.A. & Goudie, A.S. (2004) Biofilms and case hardening on sandstones from Al-Quawayra, Jordan. Earth Surface Processes and Landforms 29, 1473-1485.

Villa, N., Dorn, R. I. & Clark, J. (1995) Fine material in rock fractures: aeolian dust or weathering?  In Tchakerian, V. P. (eds.) Desert aeolian processes, Chapman and Hall, London, pp. 219-231.

von Humboldt, A. (1812) Personal Narrative of Travels to the Equinoctial Regions of America During the Years 1799-1804 V. II (Translated and Edited by T. Ross in 1907), George Bell & Sons, London.

Waragai, T. (1998) Effects of rock surface temperature on exfoliation, rock varnish, and lichens an a boulder in the Hunza Valley, Karakaram Mountains, Pakistan. Arctic and Alpine Research 30, 184-192.

Warsheid, T. (1990) Untersuchungen zur Biodeterioration von Sandsteinen unter besonderer Berücksichtigung der chemoorganotrophen Bakterien. Ph.D. Dissertation. Geomicrobiology, Universität Oldenburg, Oldenburg.

Washburn, A. L. (1969) Desert varnish.  In Washburn, A. L. (ed.) Weathering, Frost Action and Patterned Ground in the Mesters District, Northeast Greenland, Reitzels, Copenhagen, pp. 14-15.

Watchman, A. (1997) Differences of Interpretation for Foz Côa Dating Results. National Pictographic Society Newsletter 8 (1), 7.

Watchman, A. (2000) A review of the history of dating rock varnishes. Earth-Science Reviews 49,261-277.

Watchman, A. (2002) A reply to Whitley and Simon. INORA 34,11-12.

Whalley, W.B., Gellatly, A.F., Gordon, J.E. & Hansom, J.D. (1990) Ferromanganese rock varnish in North Norway: a subglacial origin. Earth Surface Processes and Landforms 15, 265-275.

White, C. H. (1924) Desert varnish. American Journal of Science 7, 413-420.

White, K. (1993) Image processing of Thematic Mapper data for discriminating piedmont surficial materials in the Tunisian Southern Atlas. International Journal of Remote Sensing 14, 961-977.

Whitley, D. S. & Annegarn, H.J. (1994) Cation-ratio dating of rock engravings from Klipfontein, Northern Cape Province, South Africa. In Dowson, T.A. & Lewis-Williams, J.D. (eds.) Contested images: Diversity in Southern African rock art research, University of Witwatersrand Press, Johannesburg, pp. 189-197.

Whitley, D. S., Dorn, R.I., Simon, H.M., Rechtman, R., & Whitley, T.K. (1999) Sally's rockshelter and the archaeology of the vision quest. Cambridge Archaeological Journal 9,221-247.

Whitley, D. S. & Simon, H.M. (2002a) Recent AMS radiocarbon rock engraving dates. INORA 32,11-16.

Whitley, D. S. & Simon, H.M. (2002b) Reply to Huyge and Watchman. INORA 34, 12-21.

Whitney, J. W. & Harrington, C.D. (1993) Relict colluvial boulder deposits as paleoclimatic indicators in the Yucca Mountain region, southern Nevada. Geological Society of America Bulletin 105,1008-1018.

Wierzchos, J., Ascaso, C., García-Sancho, L. & Green, A. (2003) Iron-rich diagenetic ninerals are biomarkers of microbial activity in Antarctic rocks. Geomicrobiology Journal 20, 15-24.

Zhang, Y., Liu, T. & Li, S. (1990) Establishment of a cation-leaching curve of rock varnish and its application to the boundary region of Gansu and Xinjiang, western China. Seismology and Geology (Beijing) 12, 251-261.

Zhou, B. G., Liu, T. & Zhang, Y. M. (2000) Rock varnish microlaminations from northern Tianshan, Xinjiang and their paleoclimatic implications. Chinese Science Bulletin 45, 372-376.