Botones de cachalote con perforación en V de Galeria da Cisterna (Sistema Kárstico de Almonda, Torres Novas, Portugal) ; Sperm-whale V-perforated buttons from Galeria da Cisterna (Almonda Karst System, Torres Novas, Portugal)

During excavation of the prehistoric necropolis of Galeria da Cisterna in 1988-1989, the remains of several individuals were recorded. The direct dating of four attests to use of the cave for funerary purposes during Beaker times. Whilst no Beaker pottery was found, a small fragment of a golden spiral and a set of V-perforated ivory buttons are diagnostic of the period. These items can be securely associated with the human remains dated to the corresponding time range. Visual inspection identified sperm whale ivory as the rawmaterial the buttons were made of, which density measurements confirmed. This indicator was chosen because it is non-destructive and discriminates well between the different types of ivory found in the Chalcolithic of Iberia. The average obtained for the 13 specimens from Galeria da Cisterna is 2.32 ± 0.12, well outside the range for elephant and hippopotamus ivories and consistent with the average value (2.2) for sperm-whale ivory published by Schuhmacher et al. (2013). The gold object was most likely made from regional sources. These data reflect on the wide intra-regional networks in operation during the Chalcolithic of the Lisbon peninsula. They also suggest that extra-regional exchange was rather more limited, as the golds of the Alentejo sites usually are of a distinct composition and sperm-whale has never been identified among their ivories (which are of North-African elephant, a rare occurrence in the Lisbon peninsula). RESUMEN


INTRODUCTION
The Galeria da Cisterna is an ancient karst outlet of the Almonda river; its entrance is located some 5 m above the extant spring. A year-round availability of water, plus the ecotonal location, help explain why the area was a focal point for human settlement throughout Prehistory (Fig. 1).
Archaeological work at the Galeria da Cisterna was first carried out between 1937 and 1942 (Paço et al. 1947;Guilaine and Veiga Ferreira 1970). In a second, 1988-1989 phase, the area affected by the older excavations was delimited and additional archaeological deposits were identified and excavated in three loci: AMD1, AMD2 and AMD3 (Maurício 1988;Zilhão et al. 1991;Zilhão and Carvalho 2011;Zilhão 1997Zilhão , 2001Zilhão , 2009Zilhão , 2021Carvalho 2007;Trinkaus et al. 2011;Martins et al. 2015). AMD1 featured a remnant Upper Paleolithic deposit under a Holocene cave earth, which formed the upper 20-40 cm of the AMD2 sequence (layer A). The latter was separated from beaver teeth-yielding, river-accumulated Pleistocene sands (layer C) by a dense lens of microfaunal remains, mostly of bats (layer B). AMD3 is a lens of organic sediments that represents a continuation of layer A of AMD2; a short stretch of outcropping bedrock separates these two loci (Fig. 2).
Layer A contained a large, chronologically heterogeneous set of finds: pottery; highly fragmented faunal and human osteological remains; stone and metal tools; personal ornaments made of shell, bone, ivory, stone, glass, and metal. The lack of internal stratigraphic differentiation is primarily due to the thinness of the deposit, compounded by animal burrowing and the impact of repeated human activity. The interior location, well beyond the twilight zone, the restricted space, unsuitable for residential purposes, and the numerous human skeletal remains suggest that AMD2 was used for funerary purposes only. Both the artefacts and the faunal remains are very probably grave goods, and the personal ornaments could relate to the finery and clothing worn by the deceased. Even though the literature tends to emphasise individual interment and collective burial in rock-cut and megalithic tombs as typical of Beaker funerary customs (Gonçalves et al. 2018), the use of natural caves as burial sites was a widespread feature of littoral-central Portugal through Prehistoric and Protohistoric times (Neolithic to Iron Age).
Most diagnostic finds from AMD2 are of unambiguous attribution to the Early Neolithic. However, later periods are also represented. For instance, diagnostic pottery and metal objects, such as wheel-made, stamp-decorated vessels, an iron earring, and a hafted iron spear show that the Galeria da Cisterna continued to be used for burial possibly until as late as the 1 st and certainly until at least the 4 th century BC (Tente and Lourenço 2016). Given the unstratified nature of the Holocene deposit, absolute ages for the accumulation were obtained via radiocarbon dating of faunal and human remains as well as of diagnostic artefacts made of organic materials. The results corroborate that AMD2 was used during the Early Neolithic, the Bell Beaker period, the Bronze Age, the Iron Age, and the Roman Era, as inferred from typological considerations.

Radiocarbon chronology
Here, our focus is on the Late Chalcolithic (Bell Beaker) context, represented by a set of V-perforated ivory buttons and a gold fragment, possibly of a spiral (Zilhão 2016). Six radiocarbon dates on human bone samples provide an age for this context (Tab. 1). As the skeletal remains were disarticulated, the minimum number of dated individuals is given by the four adult first phalanges of the right foot. Of these four, a DNA analysis was able to determine that two, F23-90 and G18-187, were of women (Olalde et al. 2018).
According to the sample significance test included in Calib 7.0.4 (Stuiver and Reimer 1993), the six dates are statistically indistinguishable: Test statistic T = 10.72; χ 2 (0.5) = 11.1 with five degrees of freedom. Together, these results constrain the burial activity to the 2nd half of the 3 rd millennium BC. In the Lisbon peninsula, available radiocarbon dates place the start of the Bell Beaker culture in the middle of the 3 rd millennium BC (Soares et al. 2007;Cardoso et al. 2013;Mataloto et al. 2013). Therefore, even though no Beaker ceramics (e.g., "maritime" vessels, Palmela bowls, or Ciempozuelos-type cooking pots) were present among the site's large pottery assemblage, we can be certain that the funerary use of Galeria da Cisterna spans a period during which these different styles of Beaker pottery decoration would seem to have co-existed (Cardoso 2014).

V-perforated buttons
Fourteen V-perforated buttons made of ivory were recovered during the excavation of AMD2, and another Trab  ; except in G20, tested to a depth of ~140 cm below surface, excavation stopped at the surface of layer C, marked by an accumulation of slabs and a dense lens of bat bones (layer B). Bottom left: The AMD2 locus seen from downstream after setting-up of the grid; the excavator is squatting in grid unit G20 (after Zilhão 2009: fig 2). Bottom middle: The AMD2 locus seen from upstream after setting-up of the grid; the excavator is squatting in grid unit G19. Bottom right: View of the stratigraphy exposed in grid unit G20 -a dark Holocene cave earth (layer A) separated from reddish fluviatile sands (layer C) by collapsed slabs associated with a lens of bat remains (layer B) (after Zilhão 2009: fig 2). In colour in the electronic version. was found in AMD1 (Fig. 3). Six are complete, three are halves, and the remaining fragments are too small and cannot be classified to type. The complete buttons correspond to the pyramidal type with square base ( The initial identification of the buttons' raw-material was based on visual observation. Some (cf. Fig.  3: 2-3, 5-6) displayed a white-over-orange/brown colouring pattern reflecting a marked separation between cementum and dentine, a feature that can be observed in the teeth of the sperm-whale. Such an identification was consistent with Schuhmacher et al.'s (2013) finding that all Portuguese V-perforated buttons up till then analysed were of sperm-whale ivory.
Several analytical techniques can be used to identify ivory. We have used the determination of the artefacts' density because this non-invasive and non-destructive approach is also among those that best discriminate between ivory types. Indeed, according to Schuhmacher et al. (2013), the average specific gravity of the ivory is 1.7 for elephant, 1.6 for mammoth, 1.8 for hippopotamus, and significantly higher, 2.2, for sperm-whale. We were able to measure the specific weight of 13 of the Galeria da Cisterna's 15 buttons. The results are given in table 2 and plotted in figure 5. Using water at its densest as a reference substance, these values, ranging between 1.91 and 2.60 g/cm³ and averaging of 2.32 ± 0.12 g/cm³, are consistent with those for the spe- cific gravity of sperm-whale ivory reported by Schuhmacher et al. (2013). These results corroborate that teeth of this marine mammal were indeed the raw-material used.

Gold
The small gold object illustrated in figure 6, probably a spiral ring fragment, comes from AMD3. Micro-EDXRF analysis of this ornament shows that it was made with a gold-silver alloy (Au 89.5 ± 0.3 wt %, Ag 10.4 ± 0.3 wt %, Cu < 0.04 wt %). This elemental composition is comparable to that of a set of gold ornaments (one double-conical bead, 11 tubular beads, two spiral beads, and one spiral ring) associated with Beaker pottery found in a collective tomb located nearby, at Convento do Carmo, in the city of Torres Novas, some 7.5 km to Southeast (Valério et al. 2017).

DISCUSSION AND CONCLUSIONS
In the Chalcolithic of Portugal, ivory finds are not infrequent. Possibly due to differential preservation issues (in the acid soils making up most of the inland and of the central and northern parts of the country, archeological contexts seldom yield organics of any kind), their distribution is largely restricted to the littoral bands of limestone substrate in the Algarve and between Coimbra and Lisbon. However, isolated occurrences exist in the interior of Alentejo. Cardoso and Schuhmacher's (2012) review names 40 sites, which yielded 163 objects, most retrieved from the differ-  ent pre-Beaker phases of the Portuguese Chalcolithic; those authors conclude that the use of ivory decreases in the Beaker period, to which no more than 40 (with a total weight of 68.3 g) of those 163 items could be securely assigned. As subsequent research has shown, these numbers incompletely reflect the importance of ivory during the 3 rd millennium BC. For the Alentejo, the incomplete nature of the picture of Copper Age ivory use available until a decade ago has been eloquently illustrated by Valera et al.'s (2015) report of an assemblage totalling 1348 fragments recovered primarily from funerary contexts within the Perdigões ditched enclosure. The majority, 837 of these fragments (1897.3 g) were found in two tholos-type collective tombs associated with the enclosure. The tombs' chronology, based on directly dated human bone remains, spans the second and third quarters of the 3 rd millennium BC. The types identified among the ivory ensemble include decorated thin plaques, cylindrical boxes, pins, bracelets, lunulae, beads, different types of idols, dagger hilts, combs, pendants, zoomorphic figurines, a staff, and buttons. The latter are represented by three specimens from Tomb 2, one assigned to its first phase of use and two assigned to the second. Among those second-phase buttons, a "little button with appendages" (Valera et al. 2015: 406) is one of the 16 ivory items from Perdigões whose provenience was determined. In all cases, the raw-material was found to come from African savannah elephants (Loxodonta africana africana).
The radiocarbon dates in table 1 are indistinguishable from those obtained at Perdigões for the second phase of Tomb 2, and for some in Assemblage 1 (Valera et al. 2015: 393-394;Valera et al. 2019: table 1.1). Despite the contemporaneity, Galeria da Cisterna's Beaker context yielded no elephant ivory; sperm-whale was the only raw-material identified among V-perforated buttons. Except for a recently reported hippo button from Convento do Carmo (Schuhmacher and Banerjee 2019), the same applies to all of the Lisbon peninsula's Beaker ivory (Schuhmacher 2017). In addition, note that the sperm-whale buttons from the tomb of Verdelha dos Ruivos (located some 20 km northeast of Lisbon) are virtually identical to those from Galeria da Cisterna, as are those from Dolmen das Conchadas (Belas, Lisbon) and the Quinta do Anjo rock-cut tombs (Palmela) illustrated by Schuhmacher et al. (2013: fig. 2: 1-6, 12). Typology and raw-material therefore concur to draw a pattern strongly suggestive of the Lisbon peninsula and the Alentejo belonging, during the Beaker period, to largely separate spheres of interaction. This conclusion is supported by the similar contrast found when comparing the elemental composition of gold objects. In the Chalcolithic of the Lisbon peninsula, they are characterized by a very low Cu content and a consistent Ag content, between 6 and 16 wt %. In the Alentejo and adjacent regions of Spain, they are characterized by a high prevalence of pure gold and a wide range of variation in Ag content (0.7 to 21 wt %) (Valério et al. 2017). These authors suggest that the contrast is due to the natural composition of the alluvial sources used in each region, and that it is indicative of "very limited trade of nuggets and jewellery with neighbouring regions."  The spring of the Almonda being located >40 km inland, its Beaker finds nonetheless imply that the Lisbon peninsula was crisscrossed by extensive trade networks. Likely, not just the gold and the sperm-whale ivory found at Galeria da Cisterna, but also other rare, valuable, and prized raw-materials and finished objects would have been distributed across, but not beyond them. Whether similarly bounded coeval networks existed in the adjacent regions of central and southern Portugal and whether the boundaries (if any) between such networks reflect the geography of the period's polities or simple source-to-destiny distance effects must remain, for the time being, an open issue.
In Chalcolithic contexts of the Lisbon peninsula for which a pre-Beaker age has been postulated, ivory provenance analyses are available for three objects only (Schuhmacher et al. 2009;Cardoso and Schuhmacher 2012): a small plaque from the hill-top settlement of Leceia (Oeiras) was made on the ivory of a fossil elephant, Elephas antiquus, while a needle from the same site and a barrel-shaped bead from one of the Quinta do Anjo tombs were made of African savannah elephant ivory. Thus, if representative, these data would seem to imply a chronological trend from a pre-Beaker/exotic (North African elephants) to a Beaker/regional (beach-stranded sperm whales) procurement.
Alternatively, we need to bear in mind that the 17 Beaker ivory objects analysed by Schuhmacher et al. (2013) plus the 15 we here report on are all ornaments whose size range is consistent with that of a sperm whale's tooth. In secure Beaker contexts of the Lisbon peninsula, larger items such as some of those makingup the Perdigões collection, ones that would have required a bulkier raw-material source, namely elephant tusk (e.g., bracelets), have yet to be analysed. In this respect, note that the study of lead isotopes suggests that the copper artefacts manufactured in the Chalcolithic of the Lisbon peninsula used Ossa-Morena Zone ores (although sources in the so-called Iberian Pyrite Belt cannot be excluded); and the sourcing of the amphibolite used through the 3 rd millennium BC to make the polished stone tools found in the region's settlement sites accords well with a similar Alentejan provenance (Cardoso and Barros e Carvalhosa 1995;Lillios 1997;Müller and Cardoso 2008;Müller and Soares 2008;Gauß 2016). Therefore, it remains possible that African elephant ivory continued to be imported and used in the region during the Beaker period, its current invisibility being simply an artefact of functional factors and research biases.
Be it as it may, enough evidence seems to exist that some of the patterns revealed by our data and review are likely real. The evidence, which we summarize in table 3, supports the following conclusions: During the Beaker period, based on the sample analysed so far, sperm-whale ivory does not seem to be making its way to such Alentejo sites as Perdigões, despite the distance involved -120 km inland from the mouth of River Sado, where the adjacent sea was used for sperm-whale hunting in the 1920s (Brito 2008)being much less than that separating such sites from the primary sources (North Africa) of the coevally used elephant ivory. This suggests that the latter arrived in the Alentejo via commercial routes linking the area to the East and South. Put another way, the direct trade connection between North Africa and the Lisbon peninsula hypothesised by Schuhmacher (2017) cannot be supported with present evidence; based on parsimony, it is most likely that the Lisbon peninsula represented instead the tail end of the routes linking the Alentejo with southern Spain (Fig. 7).
Sperm-whale would seem to be an alternative source of ivory that the Lisbon peninsula societies developed in Beaker times to supplement the North Africa-South Spain supply chain. The emergence of this alternative might reflect a growing popularity of the raw-material, implying the need to find additional (or cheaper, i.e., closer) sources to satisfy increased demand (Schumacher et al. 2013). Whether the substance's new appeal reflected a change of taste among the elites or responded to a now socially more widespread fashion must remain, however, yet another open issue. Tab. 3. Regional distribution of the ivory artefacts from the Copper Age of southern Portugal whose raw material has been determined (in parenthesis, the number that are V-perforated buttons).