The stylistically Paleolithic petroglyphs of the Côa valley (Portugal) are of Paleolithic age

A refutation of their <<direct dating>> to recent times

João Zilhão


8. Bednarik's arguments on context

Most of the contextual arguments referred to by Bednarik in his dating report to EDP (Bednarik 1995b) are essentially the same which he developed in Australia before ever having set foot in Portugal, as stated in a paper published in the April 1995 issue of the AURA Newsletter (Bednarik 1995a). He had been informed in a reply sent to him by the present author, which he received before his trip to the Côa, that these arguments were ill-founded. Notwithstanding, he used them in his report to EDP and has developed them even further in a paper to appear in a coming issue of Rock Art Research (Bednarik n. d.).

Detailed information on Portuguese Upper Paleolithic faunas can be found in Cardoso (1993). The chronological, environmental and cultural evidence for the same period has been systematically reviewed in Zilhão (1995a). Marks et al. (1994), Póvoas et al. (1992) and Zilhão (1988, 1989, 1990, 1991, 1993, 1994) contain partial discussions of the relevant data in English or French.



8.1. Frost-weathering

In the littoral region of Estremadura, due to its proximity to the at times very cold waters of the Atlantic, last glacial cryoclastism of limestones is known almost down to present sea level (Daveau 1980). But, as demonstrated by recent work on slope (Rodrigues 1991) and cave (Zilhão 1995) deposits, below 400 m this process does not seem to occur after the last glacial maximum, coming to an end with the Upper Solutrean, at the time (ca. 17,000 BP) when a very warm oscillation is recorded in sea surface temperatures off the Portuguese coast, which by then reached almost modern levels (Bard et al. 1992; Duplessy et al. 1992). Despite subsequent cold oscillations at sea, from then on the limestone hills and plateaus of this littoral region seem, on present evidence, to become covered by oak forests, and ibex and chamois disappear from the faunal assemblages. If, for the sake of argument, one used this littoral region as a model for the interior, and conceded that frost-weathering might have been a problem for the preservation of engravings of Solutrean age, the argument would still be simply untenable for the Magdalenian period. In the interior regions of North and Central Portugal, however, as a result of extreme dryness and different lithology (schists and granites instead of limestones), periglacial features such as cryoclastic slope deposits are unknown below an elevation of 700 m above modern sea level, so there is absolutely no basis to believe frost-weathering would have affected the low lying (ca. 100 m) and sheltered valley of the Côa, especially after deglaciation in the Cantabrian mountains and in the Serra da Estrela was completed, some time between 16,000 and 14,000 years ago (Turner and Hannon 1988). In any case, test excavations carried out at the newly found archaeological site of Cardina, located in the Côa valley itself (see below), have now settled the issue: here, the deposits containing a Late Gravettian (ca. 22,000 BP) archaeological context are coluvial sands entirely devoid of cryoclasts.



8.2. River erosion

Bednarik (1995a) argues that the fact petroglyphs occur right down to the floor of the Côa valley, only a few meters above water, makes it difficult to understand how Upper Paleolithic art could have survived the many fluctuations in river level that, based on evidence from other European rivers, must have happened in the Côa since the end of the Pleistocene. This is simply not a problem. Preservation of last glacial maximum archaeological remains in the flood plain of modern Portuguese rivers is demonstrated, for instance, at Terra do Manuel, where a radiocarbon dated 22,000 year old living floor located ca. 1 m below the surface was excavated in 1988-89 (Zilhão 1995a). In this part of the country, this is explained by the pattern of downcutting caused by eustatic response to lowered sea levels. Fluviatile terraces accumulated at the beginning of isotope stage 2 were thus exposed as beaches that were made available to human occupation, and were indeed occupied. Coluvial and eolian accumulation of sediments eroded from the extant slopes largely denuded of vegetation subsequently buried and protected these sites from Tardiglacial and Holocene erosion. The lowest panels at Penascosa and Ribeira de Piscos are, as far as elevation above the river is concerned, in a topographical position even more favorable, from the point of view of preservation, than that of the Late Gravettian and Proto- Solutrean habitation site of Terra do Manuel. There is therefore no reason to suggest, until detailed geological studies of the valley (as yet unavailable) eventually show otherwise, that such a position is incompatible with a Paleolithic age for the engravings. In other words, elevation above the river is, per se, a totally irrelevant issue for the argument concerning the chronology of the Côa petroglyphs.



8.3. Absence of cold adapted species

It is true that, as Bednarik (1995a) notes, cave bear, bison, mammoth, woolly rhino and reindeer are not present in the Côa art. But this is exactly what should be expected: they are not present either at any of the many Upper Paleolithic paleontological and archaeological sites that are known in Portugal and in Mediterranean Spain (Fig. 5), although Cardoso (1993) does cite one occurrence of mammoth in his list of Quaternary faunal remains from Portugal. The bone in question is a large shaft fragment classified as part of the femur of an elephantid, which he reasons should be a mammoth because other bones from the same site (Algar de João Ramos) were radiocarbon dated to ca. 14,000 BP, making an attribution to Elephas antiquus impossible. This reasoning, however, assumes that the faunal assemblage from this purely paleontological site is homogeneous and dates to a single period, which does not seem reasonable in a stratified cave environment poorly excavated in the late nineteenth century. In this context, it seems more reasonable to admit that the bone fragment in question does belong to Elephas antiquus, a species which, as Cardoso (1993) shows, survived in Portugal up to ca. 30,000 BP and could well have been represented in the otherwise quite banal Upper Pleistocene faunal assemblage from Algar de João Ramos.

All available evidence therefore suggests that the Atlantic and Mediterranean façades of Iberia, south of the Ebro, may have constituted a separate faunal province, where cold adapted species, even at the level of micromammals (Póvoas et al. 1994), did not penetrate (Aura and Villaverde 1995). Those species are also completely absent from the thousands of engraved slabs found in the Gravettian, Solutrean and Magdalenian levels of the deeply stratified cave site of Parpalló (Valencia), which were found in very rich, and radiocarbon dated, archaeological deposits, spanning the Gravettian, Solutrean and Magdalenian periods (Villaverde 1994). Horse, aurochs, red deer and ibex, plus the occasional chamois, bird or carnivore, that is, exactly the same species as those whose bones have been recovered in the archaeological deposits (Davidson 1983), are the animals represented in these works of mobiliary art (Table 1). Stylistically, such representations are also strikingly close to those found in the Côa and, given their archaeological context, are known to be of Paleolithic age beyond any reasonable doubt.

TABLE 1

Parpalló Upper Paleolithic decorated stone slabs
Animal species represented (a)

















G LS EMS LMS US SGI SGII SGIII EMA EMB UM GAL Other Total
Aurochs 2 6 6 1 3 2 1 1 2 10 7 15 3 59
Horse 1 7 14 8 12 20 6 9 9 12 15 10 5 128
Deer - 14 19 7 6 6 3 - 8 16 8 14 4 105
Ibex2 8 19 5 10 8 16 6 8 16 9 22 11 136
Other (b) - 1 4 - - - - - 1 26 2 218
Undetermined 2 27 42 29 19 19 9 14 19 29 25 51 32 320
Total 7 63 104 50 55 55 35 30 47 85 70 114 57 766


(a) after Villaverde (1994:Table 26, modified); G - Gravettian; LS - Lower Solutrean; EMS - Early Middle Solutrean; LMS - Late Middle Solutrean; US - Upper Solutrean; SGI - Solutreo- gravettian I; SGII - Solutreo- gravettian II; SGIII - Solutreo- gravettian III; EMA - Early Magdalenian A; EMB - Early Magdalenian B; UM - Upper Magdalenian; GAL - galleries (surface)
(b) chamois, fox, lynx, wild boar, wolf, mustelids and birds


Bednarik (n.d.) objects to these observations at two levels, one is factual, the other theoretical. At the factual level he restates, based on outdated (more than twenty year old) references, that cave bear is indeed present in two Portuguese faunal assemblages from the Pleistocene (those recovered at Furninha and Salemas), and that the geographical distribution of cave bear remains, which <<resembles the distribution of limestone karsts in Europe>>, indicates <<a massive taphonomic bias>>: <<the apparent absence of cave bear remains in regions lacking limestone caves (such as most parts of Portugal) tells us absolutely nothing about the former range of the species>>. This argument insists on an error of fact and entirely misses the point. The references to cave bear at Salemas are based on mistaken preliminary identifications by Zbyszewski (1963), subsequently accepted by Ferreira (1964) and Roche (1971, 1972); they have been corrected since by Torres (1979) and Cardoso (1993), who referred those remains to Ursus arctos. The bear remains recovered at Furninha have always been attributed to Ursus arctos since they were first studied by Harlé (1910-11). As a result, both Harlé (1910-11) and all subsequent authorities have given a categorical verdict on Ursus spelaeus in Portugal: the only bear species that can be recognized in the Quaternary faunal remains from the country is Ursus arctos (Torres 1979; Cardoso 1993). Therefore, Bednarik's taphonomic argument is irrelevant: the issue at stake is not why cave bear remains have not been found in non-limestone areas of Portugal, but why the species has never been found in Portuguese limestone cave deposits.

Since Cardoso (1993) reviews 21 cave sites spread all over the country, the absence of such remains is not easily explained away as due to deficient sampling. This brings up Bednarik's theoretical argument regarding cave bear and cold-adapted species: that <<absence of evidence>> does not equal <<evidence of absence>>, that is, that remains of those species may eventually be recovered in other regions of the country or in new, as yet unknown, sites. This is a quintessentially non-scientific style of reasoning, one that is commonly found in Christian fundamentalist literature under the form, for instance, of statements such as <<God exists because no one can prove that he doesn't>>. It is also a common argument in anti-evolutionist thinking, for instance under the form of statements like this: <<the fact that trilobites are absent from post-Paleozoic beds and dinosaurs are absent from pre-Mesozoic ones does not mean that their remains will not eventually be found in such deposits>>. If Bednarik were right in that criteria of absence cannot be accepted as scientific evidence, not only archaeology but also geology and paleontology would fall outside the scope of science.

The fact, however, is that, whether Bednarik likes it or not, these are well established disciplines that have developed their own scientific methodology and, in particular, have learned to deal with the issues regarding patterns of presence and absence in terms of probability statements (Dawkins 1991). For instance, given our present data base of Paleozoic and Mesozoic deposits, the probability that trilobites and dinosaurs were actually contemporaneous is so small that, for all practical purposes, such contemporaneity can be assessed as an impossibility. The data base of Portuguese Quaternary faunas is not as large as that of fossiliferous Paleozoic and Mesozoic beds all over the world, so the probability that bison, cave bear, mammoth, reindeer and woolly rhino may one day be found is not as small as in the dinosaur/trilobite example, and it cannot be considered a total impossibility, particularly in those parts of the country that are closer to the known past ranges of those species. Such a data base is, however, large enough to suggest that, if present at all, those cold-adapted species would probably have occurred only as very small and marginal populations (or even as stranded individuals) that one would not be correct in considering as part of the daily environment of Upper Paleolithic hunter-gatherers living along the western and southern shores of Iberia. The tentative (and questionable) identification of bison, megaloceros, reindeer and woolly rhino among the fine lined engravings of Siega Verde (Balbín et al. 1995), already in Spain but only some 60 km southeast of the Côa complex of art sites, might be taken, if confirmed by future research, as an indication of such infrequent occurrences. Meanwhile, <<absence of evidence>> should indeed be considered, in this case, as <<evidence of absence>>.

Índice do texto/Parte anterior/Parte seguinte/References/Figures

Início da página/Página principal/Bibliografia


Esta página foi realizada para o Nestcape Navigator 2.0
Última actualização no dia 1 de Junho 96
siljan@gemini.ci.uc.pt