Rock magnetic and petrographic characterization of prehistoric amerindian ceramics from the dos mosquises island (los Roques, Venezuela)

Myriam Rada, Vincenzo Costanzo-Álvarez, Milagrosa Aldana and Corina Campos

Myriam Andrea Rada. B.Sc. in Geophysical Engineering. Graduate Student, Universidad Simón Bolívar, (USB), Venezuela.

Vincenzo Costanzo-Álvarez. B.Sc. in Physics, USB, Venezuela. M.Sc. and Ph.D. in Geophysics, University of Toronto, Canada. Professor, USB, Venezuela. Address: Departamento de Ciencias de la Tierra, Universidad Simón Bolívar. Apartado 89000, Caracas 1086A, Venezuela. e-mail: vcosta@usb.ve

Milagrosa Aldana. B.Sc. and Doctorate in Physics, USB, Venezuela. Professor, USB, Venezuela.

Corina Campos. B.Sc. in Geological Engineering, Universidad Central de Venezuela. M.Sc. in Earth Sciences, USB, Venezuela. Professor, USB, Venezuela.

SUMMARY

Preliminary rock magnetic studies of 12 prehistoric Amerin-dian pottery shards (Dos Mosquises Island, Los Roques, Venezu-ela) have been carried out in order to characterize different clay sourcings and pottery manufaacturing techniques. Although these ceramics have been excavated in a single archeological site, sty-listic features sort them out into three distinct groups: standard-ized, heterogeneous and unknown). Petrographic analyses appear to be suiTable to distinguishing clay sources. Scatter plots of initial magnetic susceptibility (MS) versus saturation isothermal remanent magnetization (SIRM) and natural remanent magneti-zation (NRM) reveal two data trends that categorize the three sets of samples into two main groups, namely heterogeneous and standarized/unknown. The SIRM versus NRM crossplot also re-lates the three styles of ceramics with pottery shards from other islands and prehistoric Amerindian cultures. Thermomagnetic curves provide valuable information about original firing condi-tions in variable redox atmospheres.

Caracterización magnética y petrográfica de cerámicas prehistóricas amerindias de la isla dos mosquises (los Roques, Venezuela)

RESUMEN

Con el fin de caracterizar diferentes fuentes de arcilla y técnicas alfareras, se hizo un estudio magnético preliminar de 12 muestras de cerámicas prehistóricas Amerindias (Isla de Dos Mosquises, Los Roques, Venezuela). Aunque estas cerámicas se encontraron en un único sitio arqueológico, pueden ser clasificadas, de acuerdo a características estilísticas, en tres grupos diferentes (estandariza-das, heterogéneas y desconocidas). Los análisis petrográficos real-izados fueron particularmente sensibles discriminando las diferen-tes fuentes de arcilla. Los gráficos de la susceptibilidad magnética inicial (MS) en función de la magnetización remanente isotermal de saturación (SIRM) y de la magnetización remanente natural (NRM) presentan dos tendencias de datos que clasifican los tres grupos estilísticos en dos grupos principales, es decir heterogéneas y estandarizadas/desconocidas. El gráfico de SIRM en función del NRM también relaciona los tres estilos de cerámicas con muestras de cerámica de otras islas y de otras culturas prehistóricas Am-erindias. Las curvas termomagnéticas suministran información im-portante acerca de las condiciones originales de quemado en dife-rentes ambientes reductores-oxidantes.

Caracterização magnética e petrográfica de cerâmicas pré-históricas amerindias da ilha "dos mosquises" (los Roques, Venezuela)

RESUMO

A fim de caracterizar diferentes fontes de argila e técnicas da cerâmica, se fez um estudo magnético preliminar de 12 amos-tras de cerâmicas pré-históricas Ameríndias (Ilha de "Dos Mos-quises", Los Roques, Venezuela). Ainda que estas cerâmicas se encontrassem em um único lugar arqueológico, podem ser clas-sificadas, de acordo a características estilísticas, em três grupos diferentes (estandardizadas, heterogêneas e desconhecidas). As análises petrográficas realizadas foram particularmente sensí-veis discriminando as diferentes fontes de argila. Os gráficos da susceptibilidade magnética inicial (MS) em função da magneti-zação remanente isotermal de saturação (SIRM) e da magneti-zação remanente natural (NRM) apresentam duas tendências de dados que classificam os três grupos estilísticos em dois grupos principais, quer dizer heterogêneas e estandardizadas/desconhe-cidas. O gráfico de SIRM em função do NRM também relacio-na os três estilos de cerâmicas com amostras de cerâmica de outras ilhas e de outras culturas pré-históricas Ameríndias. As curvas termomagnéticas subministram informação importante sobre as condições originais de queimado em diferentes ambien-tes redutores-oxidantes.

KEYWORDS / Arqueology / Pottery / Provenance Studies / Rock Magnetism /

Received: 07/15/2007. Modified: 12/10/2007. Accepted: 01/02/2008.

 Introduction

The determination of the place where a ceramic vessel or figurine was manufactured is of major importance in archaeology, providing evidence of trade and cultural links in past times (Shepard, 1963; Orton et al., 1993).

So far, provenance studies applying rock magnetic properties have comprised measurements of the susceptibility and intensity of natural remanent magnetization of pottery samples, saturation isothermal remanence (SIRM) and anystheretic remanent magnetization (ARM) intensities (McDougall et al., 1983). These rock magnetic characterizations have been applied to obsidian samples from the Mediterranean region (McDougall et al., 1983), Argentinian and Chilean Patagonia (Vásquez et al., 2001) and central Mexico (Urrutia Fucugauchi, 1999). Most recently, CostanzoÁlvarez et al. (2006) have studied a collection of potshards from seven Venezuelan islands using a twofold magneto/dielectric technique. They concluded that: petrographic analyses agree with the clusters of data identified in scatter plots of initial magnetic susceptibility (MS) vs saturation isothermal remanent magnetization (SIRM), and natural remanent magnetization (NRM). Whereas MS, SIRMs and NRMs appear to be suiTable describing clay sources, other rock magnetic parameters, such as high temperature MS, seem to be related to different steps of pottery craftsmanship, namely clay preparation, finishing and firing. Thermomagnetic curves also provide information about original firing conditions. In the present work, as an extension of that previous research, rock magnetic studies of a set of 12 pottery shards, from a single locality, in the Dos Mosquises Island (Los Roques, Venezuela) are presented.Archeometric studies are rather incipient in Venezuela. Besides the study by Costanzo-Álvarez et al. (2006), some of the latest research in this area includes archeomagnetic studies (Brandt and Costanzo-Álvarez, 1999) and X-ray fluorescence, instrumental neutron activation analysis and prompt gamma activation analysis (Kasztovszky et al., 2004; Sajo-Bohus et al., 2005).

Studied Samples

Venezuelan prehistoric Amerindian pottery (vessels, figurines, pipes etc.) are the most important archeological vestiges of the country’s first inhabitants. In this study an attempt is made to characterize, via rock magnetism and petrographic analyses, 12 fragments of vessels found in the Dos Mosquises Island from Los Roques Archipielago (Figure 1). According to their stylistic differences these pottery shards were classified as heteroge-neous, standardized or unknown (Antczak and Antczak, 2006). The common features of the first two groups are similar to those corresponding to the Ocumaroide and Valencioide cultures, respectively. The unknown shows no resemblance with any of the other two.

Ocumaroides and Valencioides inhabited the north central part of Venezuela between 800 and 1500 AD. Venezuelan islands are arid and sandy, lacking clay deposits, sources of freshwater and suitable soils for agriculture. Thus, most of the Los Roques ar cheological sites have been interpreted as temporary camp sites established by Amerindian groups from mainland Venezuela that visited regularly this Archipielago up to the times of the European contact in the XVI century (Antczak and Mackowiack de Antczak, 1999; Antczak and Antczak, 2006).

Antropomorphic figurines are typical of the Valenciode culture that settled at the margins of the Valencia Lake (Figure 1a) as well as at the foothills of the Serranía del Interior and the Cordillera de la Costa (Cruxent and Rouse, 1958; Rouse and Cruxent, 1963). The standardized group (Figure 2a), related to the Valencioide culture, shows stylistic uniformity, elaborated shapes and colors that grade from yellowish gray to deep red. Ceramics from this group reveal a fairly advanced manufacturing technique that includes the use of reddish slip, a fine coating of clay and water applied over most of the surface. These pottery shards are also made mostly of finegrained and wellsorted clays (Antczak and Antczak, 2006).

The heterogeneous group (Figure 2b) is characterized by the rudimentary shape of the figurines and stylistic diversity that reveal a lack of advanced manufacturing techniques. Colors range from brown to reddish brown and the surface of the vessels are rough with no slip (Antczak and Antczak, 2006). The paste is coarsegrained, with large fragments of rocks and shells used for tempering. It is important to point out that no heterogeneous figurines have been found in mainland Venezuela. However, there are some vessels from archeological sites between Boca de Tacagua and Ocumare de la Costa (Figure 1a) that have similar heterogeneous stylistic features (Antczak and Antczak, 2006).

The unknown group (Figure 2c) has unique stylistic features that do not coincide with either of the other two. They reveal the use of advanced manufacture techniques similar to those that characterize the standarized group (Antczak and Antczak, 2006). Namely, thick reddish slips covering finegrained cores with colors ranging from light gray to salmon red.

Cultural strata in the site from Dos Mosquises (this study) are quite homogeneous with a thickness of ~60cm. Therefore archeological stratigraphy renders useless in establishing relative ages between different potshards. Noncalibrated ¹⁴C dates for some undifferentiated shards recovered at this site, reveal a range of ages that goes from the beginning of the XIII century up to the end of the XV century. In fact, four Dos Mosquises archeological samples excavated at 3849cm of depth reveal ages that range from a maximum of 1270 ±80 AD up to a minimum of 1480 ±80 AD (Marlena Mackowiack de Antczak, personal communication).

Several squared specimens of 1×1×0.3cm, weighing ~10g, were cut from each of the 12 shards analyzed. These squares were then embedded into a mixture of magnetically innocuous plaster of Paris and water, and let to dry for ~24h. Finally, cores of 2.5cm diameter × 2.5cm height were drilled around each specimen. These cores were used in most rock magnetic experiments. Thermomagnetic curves and thin section petrographic analyses were performed in only five shards from the standardized (2), heterogeneous (2) and unknown (1) groups.

Clay Sources and Petrographic Analyses

Clays are the chief ingredients used in pottery. They derive from the breakdown and decomposition of silicate rocks with a high percentage of alumina. Clays are either primary (residual) or secondary (sedimentary). Primary clays are those found within the same general location as the parent rock from which they originated. Residual clays contain coarse unaltered, angular fragments of their parent rocks due to incomplete alteration and decomposition of the source rocks. Secondary clays appear in contexts far removed from the parent sources. Transported clays are more common than residual ones, and because of sorting and redeposition, they usually have a more homogeneous and finer texture. Besides Si, Al and water, clays can also contain small amounts of Feminerals such as: hematite (Fe₂O₃), goethite (a-FeOOH), limonite (hydrated Fe-oxide), magnetite (Fe₃O₄), pyrite (FeS₂), marcasite (FeS₂) and siderite (FeCO₃). Grainsizes can vary between colloidal particles of 15µm to nodules of several mm in diameter.

According to Shepard (1963) each human group had a preferred location to collect their raw materials (i.e. river banks during the summer and clay quarries). In order to characterize clay sourcings for the samples of this study, thin sections from five different potshards (2 standardized, 2 heterogeneous and 1 unknown) were examined in transmitted plane (PPL) and cross-polarized (XPL) light at magnifications of 4× and 10×. No reflected ligth analyses of opaque minerals such as Fe oxydes were carried out.

All the samples analyzed have in common a quartz framework with some sand-size fragments of feldspars, pieces of granitic and metamorphic rocks, and small crystals of augite, muscovite and biotite. There are also some heavy minerals such as zircons and tourmalines. The matrix shows a mixture of sericite, biotite and Fe-oxides (Figure 3).

In most standardized samples the presence of numerous crystals of sericite and Fe-oxides indicate a superficial alteration of amphiboles, biotites and feldspars. These samples also show fragments of deformed mica and metamorphic quartz, much smaller and less angular to similar grains found in the heterogeneous samples. We argued that this material probably originated in a weathered granodiorite source as well as in the nearby metamorphic rocks from the Coastal Cordillera in north central Venezuela. On the other hand, micas and feldspars in heterogeneous samples show little evidence of weathering.

Figure 3a is a transmitted XPL photomicrograph of a standardized sample revealing the presence of a large amount of mica and plutonic quartz as well as a low percentage of plagioclase and other feldspars (more weathered). Figure 3b is a transmitted XPL photomicrograph of a heterogeneous sample showing a substantial amount of mica, low percentage of well-preserved plagioclase and other metamorphic minerals with little or no matrix. Finally, transmitted PPL thin section analyses of the unknown sample in Figure 3c reveal the presence of scattered subrounded and subangular quartz crystals of plutonic origin embedded in a massive matrix.

Pottery Craftmanship and Rock Magnetic Results

The second step in pottery craftsmanship, after collection of the clay, is to let it dry, following by grounding and compacting by water. To avoid cracks during drying, this clay and water paste is then mixed with coarser grains of groundup pottery or rocks (i.e. andesites, diorites, traquites, volcanic tuffs and ashes, sandstones and schist) altering thus the original granulometry of the raw material (i.e. temper). In the potshards from the Venezuelan islands tempering was achieved using coarse grains of ground-up seashells, sand, quartz, vegetal ashes and charcoal (Shepard 1963). After being modeled, the surface of the piece was smoothed down by hand or by rudimentary tools made of bones, wood, leather or stones. In some cases, a fine coating of clay and water was also applied to the surface of a vessel in order to produce a smooth finishing. Decorations were carved or painted with plant or mineral-based pigments. Inorganic pigments were the most commonly used by prehistoric potters. They were made mainly of hematite and limonite combined with water and an organic binder (Shepard 1963 and Rice, 1987). Depending on particle size and crystalline state of the Fe-oxides, these pigments could turn into a wide assortment of colors during firing. For instance, limonite becomes yellow, hematite turns black or red (depending on the temperatures used) whereas other hydrated Feoxides shift to red. Conversely, pigments of Mn or Mnoxides were less preferred because their limited variety of colors from brown to black (Shepard 1963, Orton et al. 1993; Rice, 1987).

Whereas petrographic analyses seem to discriminate quite well between different clay sources, it has been argued that rock magnetic parameters are perhaps more sensitive to features related to craftmanship (Costanzo-Álvarez et al., 2006).

Figure 4 shows a plot of the IRM acquisition curves (up to 400mT) for the 12 samples analyzed in this study. All these curves look almost identical revealing the dominant presence of a magnetic phase with rather low mean coercivity values between 40 and 60mT (probably Ti-magnetite).

In an attempt to better characterize the three stylistic groups recognized in these samples, we have generated a logarithmic plot (Figure 5) that shows initial magnetic susceptibility (MS) vs saturation isothermal remanent magnetization (SIRM) using a DC field of approximately 2T. All these samples show a rather wide interval of magnetic susceptibility and SIRM values spanning two orders of magnitude. Although the initial magnetic susceptibility and SIRMs are first-order measures of the amount of ferrimagnetic material in the sample, the scatter plot does not seem to clearly discriminate between the three stylistic groups identified in Dos Mosquises. In fact, this plot reveals two loosely defined data trends corresponding to the standardized and heterogeneous samples. The two experimental points corresponding to the unknown samples lie on both sides of the straight line that divides Figure 4 into the standardized and the heterogeneous data fields.

The scatter logarithmic plot of Figure 6 is similar to that of Figure 5; however, the SIRMs have been replaced by initial NRMs. Two intersecting but clearly distinct trends, seem to separate standardized/unknown ceramics from most of their heterogeneous counterparts. In Figure 6, data from this study have been also plotted against previous results from Costanzo-Álvarez et al. (2006), corresponding to different islands and archeological sites. We argue that NRMs better discriminate between different stylistic-groups, due perhaps to the fact that they are not only sensitive to concentration of ferrimagnetic materials, but also to grain size distribution, kind of magnetic minerals and intensity of the geomagnetic field at the time the pottery was fired (i.e. age of the ceramics).

Figure 6 shows that most heterogeneous samples from Dos Mosquises (this study) comprise the same range of scattered data for the easternmost island of La Blanquilla (Costanzo-Álvarez et al., 2006). La Blanquilla ceramics (ca. 1130 AD) were made by unknown Amerindian cultural groups that inhabited the westernmost part of mainland Venezuela (Marlena Mackowiack de Antczak, personal communication).

On the other hand, previous data corresponding to samples from Dos Mosquises and the nearby Archipielago Las Aves (Costanzo-Álvarez et al, 2006) share marginal values between the two data trends of Figure 5. Ceramics from Las Aves Archipielago belong to the Dabujoroide cultural group, which was contemporary to the Valencioides and Ocumaroides and inhabited the western Venezuelan coast, up to the European contact (Antczak and Antczak, 2006). The of the samples, namely standard-ized and unknown (from this study), Cayo Sal and Domusky Norte from Los Roques, Los Mangles from La Orchila and Isla del Tesoro from Las Aves (Costanzo-Álvarez et al, 2006), have a rather high and narrow range of NRM values (Figure 6). Based upon purely stylistic criteria, Cayo Sal samples from Los Roques have been classified as Valencioide (i.e. standardized) with the same age interval assigned to samples from the Dos Mosquises camp site (Antczak and Antczak, 2006). Within this data trend are also those from Domusky Norte. Although the latter is perhaps one of the oldest Ocumaroide camp sites (i.e. heterogeneous) in Los Roques (1060 to 1330 AD), the close interaction between Valencioide and Ocumaroide cultures is foremost evidenced by the presence of typical Valencioide vessels (i.e. heterogeneous) found all over this island (Antczak and Antczak, 2006). Finally, Los Mangles archeological site in La Orchila, with ages of ca. 1370 AD, has been documented as one of the easternmost allegedly Valencioide camp sites found in the Venezuelan islands (Antczak and Mackowiack de Antczak, 1999; Antczak and Antczak, 2006).

In Figure 7, intersection points from crossover plots (IRM that stands for isothermal remanent magnetization acquisition and its subsequent AF demagnetization) are shown against the outlined bounds of effective grain sizes for magnetite (Symons and Cioppa, 2000) and preliminary archeological data by Costanzo-Álvarez et al. (2006). In these experiments SIRM was imparted using a DC field of 2T. There is not an obvious definition of data clusters between heterogeneous, standardized and unknown samples in this cross plot.

Firing Atmospheres and Thermomagnetic Curves

The last step of potterymaking is the firing (Rice, 1987). Primitive potters in Venezuela did not have kilns. Instead, they used shallow pits or bonfires (i.e. open firing or clamping). Temperatures attained by open firing conditions ranged between 600 and 850ºC. The temperature of extraction from the heating sources possibly went from 125°C to 810°C (Shepard, 1963 and Rice, 1987). Most vessels show dark spots produced by irregular burning that may indicate that they were used for cooking purposes. Some others display clear surfaces with dark cores resulting from an oxidizing atmosphere that was not hot enough to burn away all of the organic constituents of the clay (Rice, 1987). During an initial (low heat) dehydration period, water is driven off to avoid rapid formation of steam and breakage. This is the reason why prehistoric potters used to maintain a slow heat at the very beginning of the firing process. Throughout the oxidation period that follows dehydration, the carbonaceous matter is burned out from the clay. Finally, at the vitrification period, the pottery constituents begin to soften and fuse together. Temperatures attained by prehistoric Venezuelan Amerindian potters were not high enough for true vitrification. However, an incipient vitrification might start between 800 and 900ºC depending on the amount and kind of impurities in the clay (Shepard, 1963, Orton et al., 1993; Rice, 1987).

In order to get some insight on the firing techniques used in the making of the heterogeneous, standardized and unknown ceramics from Dos Mosquises, magnetic susceptibility measurements were carried out in some of these samples, for continuously increasing (up to 700ºC) and decreasing temperatures. Some of these thermomagnetic cycles are shown in Figure 8, together with the second derivative used to precisely determine their Curie points. The equipment used was a Bartington MS2 with a MS2W probe that allows constant susceptibility and temperature readings of a sample inside a SM2WF furnace. The MS2W has an operational frequency of 696Hz and a precision of 4p×10^-6SI per 10^-5m³. In all these samples, Tipoor magnetite seems to be the chief magnetic mineral. This is indicated by the progressive drop of the magnetic susceptibility starting at ~400 and 450ºC.

The heterogeneous sample from Figure 8a shows a well defined Curie point at 580ºC, produced by nearly stoichiometric magnetite. This sample also changes its color to red after being heated up to 700ºC, suggesting the formation of some secondary hematite. Because susceptibility values are higher after cooling, new magnetite should have been formed as well.

On the other hand, thermomagnetic curves for a standardized and an unknown sample, shown in Figures 8b and c respectively, differ from their heterogeneous counterpart. In fact, these curves show a progressive rise in magnetic susceptibility, due to the Hopkinson peak for magnetite. Moreover, for these two samples the cooling curves run below the heating curves, thus indicating that there is no formation of new magnetic phases upon heating. Such a result could also reflect partial oxidation of Tipoor magnetite to hematite (Böhnel et al., 2002).

Previously (Costanzo-Álvarez et al., 2006) it was argued that the degree of irreversibility of a thermomagnetic curve could be used as an indicator of the extent of burning of the carbonaceous matter in a ceramic, during original firing. As a matter of fact, primitive open fires with insufficient draught create a reducing atmosphere with a short supply of oxygen that prevents the entire combustion of the fuel and favors the accumulation of reducing gases. Ceramics fired under such conditions do not obtain sources of oxygen and the carbonaceous matter is not completely burnt. Thereafter, in a thermomagnetic experiment, iron oxides will reduce to a lower state throughout initial heating, resulting into the formation of new magnetite, revealed by the cooling curve (e.g. heterogeneous sample in Figure 8a). Conversely, in open fires with a good draught, there will be enough oxygen to burn the fuel, resulting in an atmosphere where carbonaceous matter will be completely wiped out. It would be expected that such samples could have a more reversible behavior afterwards, right through a complete thermomagnetic cycle (e.g. standarized and unknown samples in Figures 8b and c).

For the complete thermomagnetic cycles of Figure 8, the likeness of these heating and cooling curves was quantified using the irreversibility parameter (IP) proposed by Böhnel et al. (2002). This IP is just the areal difference below the heating and the cooling curves, normalized by the area under the heating curve. An IP= 1 indicates maximum irreversibility whereas IP= 0 means complete reversibility. Moreover IP>0 implies that the cooling curve runs over its heating counterpart. The opposite applies for an IP<0.

The IP factors for the samples of Figure 8 are: heterogeneous= 0.36 (Figure 8a), standarized= 0.26 (Figure 8b) and unknown= 0.18 (Figure 8c). According to these results, and based on the discussion above, it can be concluded that the firing atmosphere for the heterogeneous sample was the most reducing one (i.e. less draught and therefore less burning of the carbonaceous matter). On the other hand, standardized and unknown samples were probably fired using better techniques, namely in fires with a good draught and therefore under more oxidizing conditions.

As a final remark, it can be pointed out that the IP factor, applied to the analysis of thermomagnetic curves for archeological ceramics, could be used as a cultural index or an indirect measure of how advanced the firing techniques developed by these primitive potters were.

Conclusions

From this preliminary rock magnetic study, carried out in order to characterize three stylistic pottery groups from a single archeological site in the Dos Mosquises Island (Los Roques Archipielago, Venezuela), the following conclusions can be drawn:

– Conventional, petrographic analyses appear to be the best means to distinguish between different clay sources, since magnetic information originally carried by the clay, seems to be largely obliterated, in a variable extent, by the manufacturing processes used by the primitive potters.

– From the discussion above it is also apparent that some rock magnetic parameters such as the NRM are more sensitive than others (e.g. parameters related with the effective grain sizes) when used to discriminate between different ceramic types, however none of them yields a clear cut picture.

– Thermomagnetic susceptibility curves (heating and cooling) provide valuable information about original firing conditions of the potshards.

ACKNOWLEDGMENTS

The authors thank Marlena and Andrzej Antczack (Universidad Simón Bolívar, Caracas) for generously supplying the pottery samples and relevant information about them. This research was partially funded by a PPI working fund to V. Costanzo-Álvarez and M. Aldana (Decanato de Investigación y Desarrollo, Universidad Simón Bolívar, Venezuela).

REFERENCES

1. Antczak A, Mackowiack de Antczak MM (1999) La Esfera de Interacción Valencioide. In Arroyo M, Blanco L, Wagner E (Eds.) El Arte Prehispánico de Venezuela. Fundación Galeria de Arte Nacional. Caracas, Venezuela. pp 136-154.        [ Links ]

2. Antczak MM, Antczak A (2006) Los ídolos de las islas prometidas: Arqueología prehispánica del Archipiélago de Los Roques. Equinoccio. Caracas, Venezuela. 630 pp.        [ Links ]

3. Böhnel H, McIntosh G, Sherwood G (2002) A parameter characterizing the irreversibility of thermomagnatic curves. Phys. Chem. Earth 27: 1305-1309.        [ Links ]

4. Brandt MC, Costanzo-Álvarez V (1999) A preliminary archaeomagnetic study of prehistoric Amerindian pottery from Venezuela: Interciencia 24: 293-299.        [ Links ]

5. Costanzo-Álvarez V, Suárez N, Aldana M, Hernández MC, Campos C (2006) Preliminary Dielectric and Rock Magnetic Results for a set of Prehistoric Amerindian Pottery Samples from different Venezuelan Islands. Earth Planets Space 58: 1423-1431.        [ Links ]

6. Cruxent JM, Rouse I (1958) An Archeological Chronology of Venezuela. PanAmerican Union Social Science Monographs VI. Washington, DC, USA. 277 pp.        [ Links ]

7. Kasztovszky Zs, MackowiakAntczak MM, Antczak A, Millán B, Bermúdez J, SajoBohus L (2004) Provenance study of Amerindian pottery figurines with Prompt Gamma Activation Analysis. Nucleonika 49: 107-113.        [ Links ]

8. McDougall JM, Tarling DH, Warren SE (1983) The magnetic sourcing of obsidian samples from Mediterranean and Near Eastern Source. J. Archeol. Sci. 10: 441-452.        [ Links ]

9. Orton C, Tyers P, Vince A (1993) Pottery in Archaeology. Cambridge University Press. Cambridge, UK. 269 pp.        [ Links ]

10. Rice P (1987) Pottery Analysis: a Sourcebook. Chicago University Press. Chicago, IL, USA. 584 pp.        [ Links ]

11. Rouse I, Cruxent JM (1963) Venezuelan Archaelogy. Yale University Press. New Haven, CT, USA. 170 pp.        [ Links ]

12. Sajo-Bohus L, Mackowiak-Antczak MM, Greaves ED, Antczak A, Bermúdez J, Kasztovszky Zs, Poitier T, Simonits A (2005) Incipient Archeometry in Venezuela, Provenance study of preHispanic pottery figurines. J. Radioanal. Nucl. Chem. 265: 247-256.        [ Links ]

13. Shepard AO (1963) Ceramics for the Archaeologist. Publ. Nº 609 Carnegie Institution. Washington DC, USA. 413 pp.        [ Links ]

14. Symons DTA, Cioppa MT (2000) Crossover Plots: a useful method for plotting SIRM data in paleomagnetism. Geophys. Res. Lett. 27: 1779-1782.        [ Links ]

15. Urrutia-Fucugauchi J (1999) Preliminary results of a rock-magnetic study of obsidians from central Mexico. Geofisica Internac. 38: 83-94.        [ Links ]

16. Vásquez CA, Nami HG, Rapalini AE (2001) Magnetic sourcing of obsidians in Southern South America: some successes and doubts. J. Archeol. Sci. 28: 613-618.        [ Links ]