Radiocarbon ages for the Coa petroglyphs are very similar to those obtained by Watchman (1995). Fundamental problems in the use of radiocarbon dating at Coa include evidence for the addition of younger carbon in an open system, and evidence of contamination from older sources of carbon. Radiocarbon measurements, therefore, cannot be used to decide whether the engravings are or are not of Palaeolithic age.
The Coa valley petroglyphs [ILLUSTRATION FOR FIGURE 1 OMITTED], galvanizing archaeologists world-wide (Bahn 1995b; Chippindale 1995; Clottes et al. 1995; Jorge 1995; Loendorf 1995; Zilhao 1995c; Zuchner 1995), have been a public controversy within Portugal (Salema 1995). Since their existence was made public in November 1994, the uproar created by the possible destruction of this Upper Palaeolithic-style art played an important role in stopping (at least temporarily) a major dam project that would flood the petroglyph sites (Zilhao 1995b).
There are claims that the petroglyphs are definitely not Palaeolithic. Alan Watchman argues that the engravings are younger than 1700 b.p. - based on interpretations of 14C ages associated with rock coatings (Watchman 1995; 1996). Robert Bednarik (1995a; 1995b; 1995c; 1995d) argues for ages younger than 7000 years based on microerosion dating and style. Archaeologically, a recent age for stylistically Palaeolithic art would undermine stylistic analyses of rock art and their chronological basis (Jorge 1995; Zilhao 1995a; Zuchner 1995).
My focus is whether available radiocarbon results can constrain the ages of Coa engravings.
Assumption of rock coatings as a closed system
A key assumption in radiocarbon-dating petroglyphs is that rock coatings form a closed system (Dorn 1994; Watchman 1996). The most common rock coating at Coa is silica glaze, growing on exposed rock surfaces and along subsurface joint faces. Rock coatings of mostly amorphous silica occur in warm deserts (Fisk 1971), in cold deserts (Weed & Ackert 1986), along tropical rivers (Alexandre & Lequarre 1978), and in mid-latitude humid temperate settings (Robinson & Williams 1987; 1992; (Farr & Adams 1984; Weed & Norton 1991) that has been used to radiocarbon-date rock paintings (Watchman 1992; 1994) and petroglyphs (Nobbs & Dorn 1993).
I observed two different types of silica glazes in Coa samples [ILLUSTRATION FOR FIGURE 2 OMITTED]. One is mostly silica, likely equivalent to Watchman's (1995; 1996) 'white amorphous silica' deposits. The other is silica with abundant aluminium - likely equivalent to Watchman's (1995; 1996) 'silty brown' accretion. A minor component of silica glaze, iron skin, occurs in discontinuous patches [ILLUSTRATION FOR FIGURE 2 OMITTED]. The texture and geochemistry of silica glaze on petroglyphs is similar to coatings on adjacent natural joint faces and coatings exposed in dam and road cuts.
I agree with Watchman (1995) that aluminium-rich silica glaze is more common, contains detrital grains of various minerals, and typically rests on top of aluminium-poor silica glaze. I also concur that a stratigraphy of 'silica under silty brown' (Watchman 1996: 28) is common on natural joint faces, but this stratigraphic relationship did change from cross-section to cross-section. Unlike Watchman (1996: 28), however, I did find some aluminium-poor silica coatings within petroglyph grooves [ILLUSTRATION FOR FIGURE 2 OMITTED]. I also observed a variety of complex stratigraphic superpositions of aluminium-poor, aluminium-rich and iron-rich layers.
Organic carbon was extracted from five contexts associated with Coa engravings:
i in pore spaces of rock weathering rinds on exposed panel faces;
ii in weathering rinds within petroglyph grooves;
iii in weathering rinds collected from unexposed rock crevices;
iv organic matter between rock and coating; and
v a sample from …