A Funerary Rite Study of the Phoenician–Punic Necropolis of Mount Sirai (Sardinia, Italy)

International Journal of Osteoarchaeology Int. J. Osteoarchaeol. 20: 144–157 (2010) Published online 24 December 2008 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/oa.1012 A Funerary Rite Study of the Phoenician–Punic Necropolis of Mount Sirai (Sardinia, Italy)

G. PIGA,a* M. GUIRGUIS,b P. BARTOLONI,b A. MALGOSAa AND S. ENZOc a Anthropology Unit, Department of Animal Biology, Vegetal Biology and Ecology, Universitat Auto`noma de Barcelona, 08193 Bellaterra, Spain b Department of History, Viale Umberto 52, 07100 Sassari, Italy c Department of Chemistry, Via Vienna 2, 07100 Sassari, Italy

ABSTRACT A recent excavation in the Phoenician–Punic necropolis of Mount Sirai, located in the south- western part of Sardinia, Italy, has brought to light a number of tombs contextually attributed to a period from the early 6th to early 5th century BC, which is simultaneous with the beginning of the Carthago influence in Sardinia. Among the interred burials recently brought to light, the skeletal remains, sometimes of two superposed bodies, are found in a primary position and with fine anatomic connection. Some of the bones were visually stained, suggesting they were possibly subjected to fire treatment. In order to ascertain more objectively whether the bodies were subjected to burning, the bones from all the tombs were investigated by powder X-ray diffraction (XRD) and Fourier Transform infra-red (FT-IR) spectroscopy techniques. After excluding the role of important diagenetic effects, from line broadening/sharpening analysis of hydroxylapatite in the bones according to the Rietveld method, it was evaluated that the bodies were probably subjected to a temperature regime from 300 to 7008C. These data were supplemented and confirmed by an analysis of the splitting factor (SF) of apatite phosphate peaks in the infra-red spectrum of the bones. Our results indicate the existence of a rite intermediate between incineration and inhumation. This sort of ‘semi-combustion’, perhaps limited to the period of the early 5th century BC, appears to be peculiar just to this site. Copyright ß 2008 John Wiley & Sons, Ltd.

Key words: Phoenician–Punic necropolis; burned bones; powder X-ray diffraction (XRD); infra-red spectroscopy (FT-IR); funerary rite

Introduction peopling from an anonymous settlement close to the actual village of Portoscuso was also Historical background to the site suggested, but appears to be unlikely. It is documented that around the year 540 BC, The site of Mount Sirai, located in the south- Carthago decided to subject the island to military western part of Sardinia near the city of Carbonia, occupation, but a coalition of Phoenician cities in was established around 740 BC and very likely Sardinia, certainly involving Sulcis and Mount inhabited by the Phoenicians from the nearby Sirai, ﬁrmly resisted this expansion. However, a city of Sulcis (today known as S. Antioco). Its few years later Carthago organised a second military expedition that defeated the Phoenician alliance. The population of Mount Sirai was * Correspondence to: Anthropology Unit, Department of Animal massacred and the city almost completely Biology, Vegetal Biology and Ecology, Universitat Auto`noma de Barcelona, 08193 Bellaterra, Spain. destroyed. It is estimated that after this event e-mail: [email protected] only a dozen families were inhabiting the village.

Copyright # 2008 John Wiley & Sons, Ltd. Received 28 January 2008 Revised 25 March 2008 Accepted 2 June 2008 Funerary Rites at a Phoenician-Punic Necropolis 145

This situation remained approximately the same repopulation of the site by new colonists, the until 360 BC, when Carthago decided to funerary rite changed suddenly. Thus, incinera- strengthen various Sardinian sites, including tion was replaced by inhumation, according to Mount Sirai. what was well-established in Carthago as well as After the year 238 BC, coincident with the amongst the northern African populations (Bar- change of domination from Carthago to Rome, toloni, 1981, 1996). In this respect, it is also the so-called neo-Punic period, the fortress of worth noting another Punic custom, consisting of Mount Sirai was completely demolished. A new the storage of infant bodies inside transportation city plan included four large building arrays. amphora (‘enkythrismo`s’). Around 110 BC, the inhabitants of Mount Sirai, probably because they inhabited a naturally well- defended hill, were deported by Rome within a General considerations about the recent framework of repression of frequent insurrec- Mount Sirai tombs tional riots. So the site was abandoned and only inhabited sporadically (Bartoloni, 2000). The most recent excavations of the site have An initial excavation was carried out between brought to light 18 tombs contextually attributed 1963 and 1966 AD (Barreca, 1964; Amadasi & to a period from the early 6th to the early 5th Brancoli, 1965; Amadasi, 1966, 1967). After a century BC, which coincides with the beginning break during the 1970s, investigations were of the Carthago influence in Sardinia. resumed in 1980 by the groups led by P. In this type of interred burial the skeletal Bartoloni and M. Botto (Bartoloni, 2000; Botto remains, sometimes of two superposed bodies & Salvadei, 2005). These systematic excavation (see Figure 1), were discovered in a primary campaigns originally involving a large graveyard position and with very good anatomical connec- area were further continued between 2005 and tion (see Figure 2a and 2b). Note also that tomb 2007 (Guirguis, 2006). 16 was actually an ‘enkytrismo`s’ with a few infant Overall, it was established that the Phoenicians remains inside (Figure 3). adopted mostly incineration rites for their dead, similar to other Phoenician cities of Sardinia (Nora, Tharros, Portoscuso, Bitia and Pani Loriga) and of other territories of the western Mediterranean Sea (Bartoloni, 1981, 1990, 1995, 1996, 2003; Bartoloni & Tronchetti, 1981) In the years between 1981 and 1987, 75 ground tombs of Phoenician attribution were dated between the late 7h century and ca. 525 .C, when the incineration rites were mainly practiced. Nevertheless, cases of inhumation were also coexisting. This may have derived from the local population, enforced by subsequent Carthago influence. As a matter of fact, it seems likely that the inhumation rite survived during and after the Carthaginiensis period because it was previously practiced by the inhabitants of Nuragic origin who contributed, together with the Phoenicians, to the population of the first urban nuclei of Sardinia (Barreca, 1985a,b; Bartoloni, 1985; Ugas & Lucia, 1987). It is also accepted that, after the conquest of Figure 1. Tomb 8 with two overlapped bodies. This figure Mount Sirai by the Punics from Carthago dated is available in colour online at www.interscience.wiley. 525 BC (Bartoloni et al., 1997), and during later com/journal/oa.

Figure 2. The bodies of (a) tomb 6 and (b) tomb 12 respectively, from where the primary position and good anatomical connection can be appreciated. This ﬁgure is available in colour online at www.interscience.wiley.com/journal/oa.

Figure 3. Tomb 16 consists of a transportation amphora (‘enkythrismo`s’) containing the remains of a child. This ﬁgure is available in colour online at www.interscience.wiley.com/journal/oa.

Copyright # 2008 John Wiley & Sons, Ltd. Int. J. Osteoarchaeol. 20: 144–157 (2010) Funerary Rites at a Phoenician-Punic Necropolis 147

Figure 4. A femur detail from Tomb 8 where the brown dark colour is evident, and is probably derived from ﬁring. This ﬁgure is available in colour online at www.interscience.wiley.com/journal/oa.

In some bodies a dark-brown colour was growth of the average crystallite size of hydro- observed on the bones that may be attributed to a xylapatite (HA), which can be measured rela- burning process (see Figure 4). Numerous post- tively well by the line broadening/sharpening depositional and post-burial agents affect and analysis of diffraction peaks. potentially alter skeletal material (Gifford, 1981). In the first critical study of its kind, Shipman However, it is often difficult to interpret such et al. (1984) investigated the microscopic agents due, in part, to the absence of known morphology of various osteological materials signatures of certain taphonomic processes. and used X-ray diffraction in order to assess In order to ascertain using objective tools whether specimens of unknown taphonomic whether all the bodies were subjected to burning history were burnt, and the maximum tempera- and to what extent, the bones recovered from ture reached by those specimens. Like the tombs were investigated by both powder X-ray previously cited studies, these investigations were diffraction (XRD) and Fourier Transform infra- based on the fact that heating bones causes a red (FT-IR) techniques which, under specific sharpening of diffraction patterns, attributed to assumptions, have been demonstrated to be able increased crystallite size and decreased lattice to discriminate the degree of fire treatment to strain of osteological phases. which bones were possibly subjected (Shipman For simple evaluations, the line-broadening et al., 1984; Shahack-Gross et al., 1997; Enzo et al., analysis of XRD patterns may be applied just to 2007). the (002) reflection at half height of hexagonal The XRD technique was first applied to apatite (Sillen, 1989; Tuross et al., 1989; Hedges archaeological subjects in 1964 (Perinet, 1964). et al., 1995; Hedges & Millard, 1995). The limits Later Bonucci and Graziani (1975) demonstrated of a single line profile approach for determining that high temperatures of fire treatment induce a average crystallite size are well known (Wagner,

1966). In fact, more complex crystallinity indices Thorp & Van der Merwe, 1991; Rink & Schwarcz, have been developed, although in a limited 1995; Sillen & Sealy, 1995; Stiner et al., 1995; angular range (e.g. 30–408 in 2u using CuKa Sillen & Parkington, 1996; Wright & Schwarcz, radiation) (Bartsiokas & Middleton, 1992; Person 1996; Shahack-Gross et al., 1997). The phosphate et al., 1995). and carbonate absorption bands of apatite are Since the overall XRD peak broadening is generally constituted by overlapped lines whose actually related to the inverse of the domain size width decreases as a function of treatment extension and to the degree of lattice disorder temperature, according to an empirical crystal- (also referred to as microstrain), we have recently lisation index also called splitting factor (SF). In updated the X-ray line-broadening approach by analogy to the work of Surovell and Stiner (2001), collecting patterns in an extended angular range we have produced a calibration for the splitting (158–1408 in 2u), employing long times of factor SF (Weiner & Bar-Yosef, 1990) as a acquisition and using the Rietveld reﬁnement function of selected temperatures using a recent method (Piga et al., 2007a,b, 2008), in analogy human bone. with the approach suggested by Michel et al. The calibration procedure for IR bands has (1995). This method is supposed to be more the same grounds as that used to evaluate the reliable for a precise description of the growth broadening/sharpening effects observed in the phenomena induced in the hydroxylapatite (HA) XRD pattern, although the basics of interaction micro-(or nano-)crystals as a function of ﬁre between light radiation and matter in the two temperature, since it is not limited to the analysis techniques here considered are completely of a few selected peaks, but evaluates the different. broadening of all diffraction peaks collected. To further support the XRD results, the same bone specimens have been investigated by FT-IR Materials and methods spectroscopy, whose absorption bands are related to the bond strength of carbonate and phosphate Table 1 reports a list of bone specimens from the groups of apatite (Weiner & Bar-Yosef, 1990; Lee- 18 tombs of the Mount Sirai necropolis which

Table 1. List of the studied bones and the chronology attributed according to the tomb contexts

Sample code Part of the body examined Chronology and period

Tomb 248 Femur Beginning 6th (590–570 BC) Phoenician Tomb 253 Femur Beginning 6th (560–540 BC) Phoenician Tomb 245 Femur Middle 6th (550–530 BC) Phoenician Tomb 256 Tibia Middle 6th (560–540 BC) Phoenician Tomb 257 Femur Middle 6th (570–540 BC) Phoenician Tomb 255 Infantil fragment Second half 6th (550–520 BC) Phoenician Tomb 7 Femur Second half 6th (540–510 BC) Tomb 3-1 Pediphalanx End 6th Tomb 236 Homerus End 6th–start 5th (510–490 BC) Punic-Phoenician Tomb 3-2 Homerus End 6th–start 5th (520–480 BC) Punic-Phoenician Tomb 5 Homerus End 6th–start 5th (520–480 BC) Punic-Phoenician Tomb 6 Fibula End 6th–start 5th (520–490 BC) Punic-Phoenician Tomb 14 D Sacrum End 6th–start 5th (520–490 BC) Punic-Phoenician Tomb 14 B Cranium End 6th–start 5th (520–490 BC) Punic-Phoenician Tomb 15 Fibula End 6th–start 5th (520–490 BC) Punic-Phoenician Tomb 8-1 Femur Start 5th (500–480 BC) beginning Punic Tomb 8-2 Rib Start 5th (500–480 BC) beginning Punic Tomb 10 Fibula Uncertain; probably start 5th (500–480 BC) beginning Punic Tomb 12 Femur Start 5th (500–480 BC) beginning Punic Tomb 13 Femur Uncertain; probably start 5th (500–480 BC) beginning Punic Tomb 16 Femur First half 5th (490–450 BC) Punic

Copyright # 2008 John Wiley & Sons, Ltd. Int. J. Osteoarchaeol. 20: 144–157 (2010) Funerary Rites at a Phoenician-Punic Necropolis 149 were examined by XRD and FT-IR. According to incorporated in the Origin1 software, assuming the context, the tombs are attributed to a period a polynomial background of order 1 or 2 and encompassing the end of 6th to the beginning of symmetric Pearson VII type functions for the the 5th century BC. transmitted line shape. An advantage of using Samples of c. 0.5 g were prepared for XRD by Pearson VII functions is the possibility of hand-grinding with an agate mortar and pestle accounting for a shape parameter m changing until reduced to a sufficiently fine powder. The continuously from Gauss (m ¼1) to Cauchy X-ray diffraction patterns were recorded over- (m ¼ 1) and even to super-Lorentzian (1 > m > 0) night using Bruker D8, Philips PW-1050, Sie- according to the nature of the absorbing mens D-500 and Rigaku D/MAX diffractometers processes involved (Enzo & Parrish, 1983). in the Bragg–Brentano geometry with CuKa The human samples from Sassari ossuary radiation (l ¼ 1.54178 A˚ ). The goniometer graveyard used for FT-IR calibration were was equipped with a graphite monochromator heat-treated with a heating rate of 208C/min at in the diffracted beam and the patterns were selected temperatures (200–10008C) in air using collected with 0.058 step size. The X-ray a NEY muffle furnace. In particular, the cluster generator worked at a power of 40 kV and of bands of HA in the range 500–700 cm1 30mA, and the resolution of the instruments is analysed because it is generally recognised (divergent and antiscatter slits of 0.58)was as the most reliable zone to define the splitt- determined using a-SiO2 and a-Al2O3 standards ing factor SF as a function of temperature free from the effects of reduced crystallite treatment. size and lattice defects. The powder patterns were collected in the angular range 158–1408 in 2u, with a counting time of 40 sec per point, and were analysed according to the Rietveld Results and discussion method (Rietveld, 1967) using the programme MAUD (Lutterotti et al., 1998). This is an FT-IR calibration efficient approach that evaluates quantitatively the amount, structure and microstructure We have produced a calibration of the FT-IR parameters of mineralogical phases while also band sharpening after treating a sample of human taking into account the instrumental parameters. bone at various temperatures. For a quick and While the average crystallite size parameter (also reliable measure of the FT-IR sharpening, Stiner referred to as domain size) does not depend on et al. (1995), after making reference to the 1 the order of reflections, the lattice strain does. n4(PO4) groups from 500 to 700 cm , defined The program calculates numerically the con- the quantity SF as the sum of intensities at volution equations in order to correctly dis- 565 and 600 cm1, divided by the intensity of the tinguish and evaluate both terms in the valley between them. This requires a careful and experimental peak broadening. Also, one import- objective determination of the background, ant advantage of the Rietveld method is that no which might be affected by different absorb- standard is required for quantitative evaluation of ing/emitting energy modes according to the phases, thus minimising the work on sample chemical composition or thermal treatment to preparation. which bones were subjected. In the absence of The ground powders were mixed with KBr in overlapping bands, one may reasonably trace the weight ratio 1:100 respectively to make a straight line connecting the two minima pellets suitable for FT-IR spectra, which were from the right to the left-hand side of the collected with a JASCO FT 480 spectrometer in cluster. An alternative, more elegant method can terms of absorbance vs wavenumber n in the make use of a non-linear least-squares approach range 400–4500 cm1. Particularly, the data according to Michel et al. (1996) that we have for the phosphate band n4(PO4) in the range applied to our data for an untreated bone using 500–700 cm1 were processed using standard symmetrical Pearson VII functions as shown in non-linear least-squares fitting procedures Figure 5.

Pearson VII functions is in excellent agreement with results obtained by Michel et al. (1995) and suggests that the background may lie below what is instinctively expected. In particular, the parameters of the linear background are correlated to the shape character 1000 m of the Pearson VII functions. In turn, a super- Lorentzian character, which seems to be the 900 case with the line at ca. 560 cm1, may be related to a wide or multimodal distribution of active 800 IR states. Nevertheless, to be homogeneous 700 with similar studies so far reported in the literature, we have proceeded according to the 600

Transmittance simpliﬁed approach illustrated by Surovell & 500 Stiner (2001). In Table 2 we report the experimental values 400 for SF as a function of treatment temperature. The 300 plot of the SF as a function of applied temperature 200 (data points) shows a logistic behaviour (see Figure 7), in analogy with the calibration of XRD data, that, however, was involving the growth of 500 550 600 650 700 the HA microcrystals and not the frequency of

-1 the active bands. Even in this case the logistic Wavenumber / cm function was fitted to the data and referred to as a Figure 5. The FT-IR transmittance peaks of the phosphate calibration curve. It is possible to argue that the group where the SF is evaluable. Experiment are data FT-IR information occurs in the same temperature points, the full line is the result of a numerical fitting range as XRD (i.e. the SF increases for process employing four Pearson VII functions. Note the appearance of a shoulder at 634 cm1 for temperatures temperatures higher than 6008C and the process above 7008C, as indicated by the arrow. This figure is appears complete at ca. 10008C). available in colour online at www.interscience.wiley.com/ Sillen and Stiner (2001) reported slightly journal/oa. different data for similar heat treatments. In particular the SF values were similar to ours for temperatures below 6008C, but increased suddenly at ca. 7508C to decrease slightly to values The data relevant to our calibration at the around 6.5 at high temperatures. quoted temperatures are presented in Figure 6. Various reasons can be advanced to explain the We can observe that the two main bands of the discrepancies in FT-IR SF values from different peak cluster at ca. 565 cm1 and 600 cm1 laboratories. As already pointed out, among the respectively, becoming sharper as the tempera- various sources of subjectivity in evaluation of SF, ture increases. Also, as this sharpening proceeds, the choice of the background appears crucial. In for temperatures between ca. 700 and 8008Ca fact, if we examine our cluster of bands at high further band emerges at 634 cm1 and persists temperatures, we may suspect for the background until 10008C. behaviour that a parabola may be more suitable One sees that the data points of the than a linear trend. In both cases the experimental experimental spectrum can be satisfactorily data are very well reproduced even with accounted for by a cluster of four bands (full sophisticated methods, but it is clear from a lines) because of the presence of one shoulder at comparison of the numerical analyses reported in ca. 573 cm1 and one more weak and broadened Figure 8a and 8b that the values of SF may at ca. 635 cm1 respectively. The decomposition oscillate. However, the subjectivity is sensibly of the cluster in terms of individual best-fit reduced when dealing with low-temperature

Copyright # 2008 John Wiley & Sons, Ltd. Int. J. Osteoarchaeol. 20: 144–157 (2010) Funerary Rites at a Phoenician-Punic Necropolis 151 Transmittance / arb. un

500 550 600 650 700 Wavenumber / cm-1

Figure 6. The numerical peak decomposition operated in the as-received bone used to carry out the calibration procedure for temperature. This ﬁgure is available in colour online at www.interscience.wiley.com/journal/oa.

treatments because the intensity levels of the XRD and FT-IR analysis on Mount Sirai valley are quite insensible to the alternative remains choices of background. Keeping in mind these general limitations, In Figure 9 we show some experimental diffrac- which to a certain extent are common to all tion patterns (data points) and the relevant experimental techniques, we have proceeded to Rietveld fit (full lines) for some osseous specimens estimate the temperatures from the FT-IR SFs that are of peculiar significance for the Mount measured in the bones of Mount Sirai tombs. Sirai necropolis context brought to light in the recent excavation. It is worth remembering that the sequence of X-ray peaks represents an overlapping fingerprint of the mineralogical Table 2. The data of FT-IR SF from the phosphate bands of apatite as a function of treatment temperature phases of the specimen. These phases were easily retrieved according to an automatic search-match Temperature 8C Splitting Factor SF procedure. Subsequently, the nature of phases is confirmed with the Rietveld fit and evaluated Not burned 3.1 200 3.16 quantitatively after the refinement stage. 300 3.23 As expected, the XRD analysis points to the 400 3.26 presence of HA as the main mineralogical phase, 500 3.39 600 3.73 which is accompanied by varying quantities of 700 4.09 calcite (CaCO3) up to a maximum level of 800 5.04 31 wt.%. Sometimes, small quantities of quartz 900 6.82 1000 6.94 and clay minerals were also observed at the limits of the detection range.

7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 Splitting factor / arb. un. 3.0 200 400 600 800 1000 Temperature / °C

Figure 7. FT-IR SF behaviour as a function of temperature treatment (data points) and the best-ﬁt logistic function (full line) used for calibration. This ﬁgure is available in colour online at www.interscience.wiley.com/journal/oa.

At the moment we attribute calcite to an ground that was ﬁlling the excavated exogenous origin with respect to the osseous sepulchres, capped on top by ﬂat stones. It is material, since the presence of carbonate likely that the ground penetrated into the burial 2 units CO3 that may substitute for phosphate through the interstices of the top due to 3 groups PO4 in the apatite structure, and that can weathering effects. The fact that small be separated during the deposition times of the amphorae and other ceramics with narrow bones, may amount to not more than 7–8 wt.% necks were discovered empty inside, further (Wopenka & Pasteris, 2005). supports this hypothesis. The varying amount of calcite found in some Table 3 collects the average domain size of HA Mount Sirai bones may be related to the tufa microcrystals (or crystallites) after separating the Transmittance / arb. un arb. / Transmittance Transmittance / arb. un

500 550 600 650 700 500 550 600 650 700 Wavenumber / cm-1 Wavenumber / cm-1

Figure 8. The data for a bone specimen heat-treated at 10008C (data points) may be ﬁtted equally well whether considering (a) parabolic or (b) linear background behaviour, giving rise to some arbitrariness in the evaluation of SF. This ﬁgure is available in colour online at www.interscience.wiley.com/journal/oa.

Copyright # 2008 John Wiley & Sons, Ltd. Int. J. Osteoarchaeol. 20: 144–157 (2010) Funerary Rites at a Phoenician-Punic Necropolis 153

T15 T15

T13 T13

T16 T16

T12 T12

T7 T7

T6 T6

T3 T3

T8 T8

X-ray Intensity x Q / arb. un. T10 X-ray Intensity x Q / arb. un. T10

T5 T5

20 40 60 80 100 120 30 33 36 Scattering Angle 2θ Scattering Angle 2 θ

Figure 9. XRD patterns of some bones from the Monte Sirai necropolis. In (b) the patterns are restricted to the range from 30 to 398 in 2u, where originally a Crystallisation Index was defined. This entire collection of XRD patterns points out the important approximations assumed in determining the microstructural properties of Hydroxylapatite (HA) just from one or few selected peak profiles. With the Rietveld method the goodness of fit between the calculated and experimental pattern is measured in terms of numerical agreement factors, so this approach appears the most complete for evaluating experimental data quality (i.e. signal-to-noise-ratio) and/or the credibility of model assumptions simul- taneously. This figure is available in colour online at www.interscience.wiley.com/journal/oa.

Table 3. XRD average crystallite size, their estimated temperature, calcite wt.%, FT-IR SF, and its estimated temperature for the bones considered

Sample Average Estimated Wt.% calcite Splitting Factor (SF) Estimated code crystallite temperature/8C calculated/0.05 temperature/8C with size (A˚ )/5 with XRD technique FT-IR technique

Tomb 248 224 600 0 3.53 590 Tomb 253 264 <750 0 4.45 710 Tomb 245 241 650 0 3.83 639 Tomb 256 290 ffi700 0 4.26 690 Tomb 257 240 650 0 4.2 670 Tomb 255 205 400 31 3.27 410 Tomb 7 233 ffi650 <1 4.23 680 Tomb 3-1 252 650 4 3.67 620 Tomb 236 245 650 0 4.21 660 Tomb 3-2 187 ffi300 29 3.25 390 Tomb 5 260 650 3 5 760 Tomb 6 250 650 14 3.61 605 Tomb 14 D 220 600 < T < 700 0 3.59 610 Tomb 14 B 222 600 < T < 700 0 3.71 625 Tomb 15 172 Not burned 3 2.99 Not burned Tomb 8-1 251 650 5 4.2 670 Tomb 8-2 248 650 12 4.45 710 Tomb 10 258 650 15 3.96 650 Tomb 12 220 600 < T < 700 20 3.59 610 Tomb 13 202 500 14 3.97 650 Tomb 16 218 ffi650 5 3.55 600

component due to lattice strain from broadening dedicated to funeral pyres), discovered nearby (column 2), the relevant temperatures to which the grave area during the recent excavations, and the bones were probably subjected according to later transferred and deposed into the burials our recent studies (Piga et al., 2007b, 2008) together with the funerary items. However, (column 3) and the percentage of calcite (column preservation of the anatomic connection during 4). The temperature values retrieved from the SF transportation appears difficult to explain, unless of FT-IR bands are also reported in column 5 of the combustion process occurred with limited Table 3, and appear to be in overall agreement time of residence at the maximum fire intensity. with the XRD determinations, except for tombs As a matter of fact, the ‘pugilistic attitude’ 3-1, 5 and 13, respectively. The temperatures (Knight, 1996) observed also by Bohnert et al. estimated by FT-IR (column 7) appear to be (1998) is absent in the bodies of Mount Sirai, slightly higher than the corresponding values where arms and phalanxes are in a perfect supine obtained by XRD. position. It should also be mentioned that Overall, all bodies seem to have been fired to Bohnert observed the pugilistic attitude for ca. temperatures not higher than 7008C. In the case 10 min after flame spray fired the body at a of Tomb 13 we determined a fire temperature of constant temperature of 7208C. However, this 3008C, but this value has a relatively high situation appears to be quite different from what associated variance because below 5008C the occurs in a real process of cremation. Never- XRD line-broadening may not be very precise. In theless, it is still possible that an intense fire fact, both techniques may not be reliable to assess carried out in the ‘ustrinum’ for a short time did cremations carried out at temperatures lower than not completely destroy the bodies, allowing their 5008C. In any case, the FT-IR and XRD data subsequent transportation to the tombs. confirm unambiguously that all the bodies In any case, it is possible that the typical rite brought to light in the recent Mount Sirai documented here was simply aiming at eliminat- excavation (except for Tomb 15) were burnt. ing the fleshy parts of the bodies. If this is the The specimen from tomb 15 shows an average case, we could identify the rite as a hygienic/ domain size of 172 A˚ , in close agreement with cleaning process, maybe adopted for deaths due values determined in inhumated bones (Piga et al., to contagious diseases and/or infection pathol- 2007b, 2008). This observation allows us to ogies. exclude an important role for diagenetic effects in However, even the simple hygienic motivation the sharpening of XRD lines, as was advanced by cannot be entirely convincing. The care given to Sillen and Le Gros (1991) in the case of ancient the bodies, the valuable items found in the tombs bodies buried in acid grounds. Actually, the and the persistent adoption of this practice across limestone-based (calcite) ground of our burials the ages suggest that the ‘semi-combustion’ should maintain a chemically alkaline character process also had a symbolic motivation, probably to the local environment (pH 8.2), which related to faiths well-established in the com- should preserve the microstructure of bones munity of the necropolis. rather than deteriorating it in terms of apparent We may hypothesise that between the 6th growth of crystallites, with consequent unreliable and 5th centuries BC at the start of the Punic estimates for fire temperature (Hare, 1980). domination, part of the population of Mount The precise modalities according to which the Sirai, related to the original Phoenician com- bodies were fired are not totally clear, as neither munity from near-east Mediterranean, strongly traces of combustion nor charcoal or wood were maintained the funerary practice of cremation. recognised at the bottom or in the walls of burials, This is demonstrated by the temporal evolution in or in the funerary ceramic miscellanea. Because of the necropolis of the 6th century BC that this absence it seems that the bodies were fired evidenced the coexistence of a row of tombs first and only later were the items buried in the with hypogeal rooms and of another row with tomb. ground burials. An alternative possibility is that the bodies Other literature on Phoenician and Punic were first fired in the ‘ustrinum’ (i.e the site customs (Beńichou-Safar, 1982; Rodero Riaza,

Copyright # 2008 John Wiley & Sons, Ltd. Int. J. Osteoarchaeol. 20: 144–157 (2010) Funerary Rites at a Phoenician-Punic Necropolis 155

2001) does not report observations similar to the References case of Mount Sirai here discussed. Amadasi MG. 1966. L’abitato, Monte Sirai – III. Rapporto preliminare della campagna di scavi 1965. Studi Semitici 20: 83–103. Conclusion Amadasi MG. 1967. La zona C, Monte Sirai – IV. Rapporto preliminare della campagna di scavi 1966. Recent excavations still continuing at Mount Sirai Studi Semitici 25: 55–93. have brought to light 18 tombs, contextually Amadasi MG, Brancoli I. 1965. La necropoli, Monte attributed to a period from the early 6th to the Sirai – II. Rapporto preliminare della campagna di th scavi 1964. Studi Semitici 14: 95–121. early 5 century BC, showing the skeletons in a Barreca F. 1964. Gli scavi, Monte Sirai – I. Rapporto primary position and fine anatomic connection. preliminare della campagna di scavi 1963. Studi A comparative analysis using powder X-ray Semitici 11: 21–36. diffraction and Fourier Transform Infra-red Barreca F. 1985a. Fenici e Cartaginesi in Italia. spectroscopy established that all the bodies L’archeologia fenicio-punica in Sardegna. Bollettino examined, except for Tomb 15, were fired before d’Arte 31–32: 57–95. burial in a temperature range from 300 to 7008C. Barreca F. 1985b. Sardegna nuragica e mondo The issues stimulated by this result have been fenicio-punico, Civilta` nuragica, Milano Eds, p. 317. essentially developed for understanding the Bartoloni P. 1981. Contributo alla cronologia delle technical practice of the firing process and the necropoli fenicie e puniche di Sardegna. Rivista di anthropological meaning of the rite. Studi Fenici 9 (suppl.): 13–29. Bartoloni P. 1985. Monte Sirai. 1984. la necropoli This investigation has allowed assignment of (campagne 1983 e 1984). Rivista di Studi Fenici 13: the sepulchres with time continuity from the 6th th 247–263. to the 5 century BC, and allowed us to ascertain Bartoloni P. 1990. Riti funerari fenici e punici nel to a very fine level of detail the conversion from Sulcis. Quaderni della Soprintendenza Archeologica per le incineration rite to inhumation, demonstrating Province di Cagliari e Oristano 6 (suppl.): 67–81. the practice of an intermediate process of ‘semi- Bartoloni P. 1995. L’insediamento di Monte Sirai nel combustion’. quadro della Sardegna fenicia e punica: AA.VV., It will be a challenge for future work extended Actes du IIIe Congre`s International d’Etudes pheńiciennes et to other Sardinian sites to assess whether the data puniques, Tunis, 11–16 November 1991: 99–108. collected at Mount Sirai corresponds to a Bartoloni P. 1996. La necropoli di Bitia. Collezione di situation common to other Sardinian cities, or Studi Fenici 38. Bartoloni P. 2000. La necropoli di Monte Sirai. rather represents the result of a local and peculiar Col- lezione di Studi Fenici 41. culture, possibly typical of the Sulcis territory, Bartoloni P. 2003. Le necropoli fenicie di Sardegna: being so dense with Phoenician–Punic locations. Fra Cartagine e Roma II. Seminario di studi italo- tunisino. Epigrafia e Antichita` 20: 57–69. Bartoloni P, Tronchetti C. 1981. La necropoli di Nora. Collezione di Studi Fenici 12. Acknowledgements Bartoloni P, Bondı` SF, Moscati S. 1997. La penetra- zione fenicia e punica in Sardegna. Trent’anni dopo. G. Piga thanks Prof. Vittorio Mazzarello, Prof. Memorie dell’Accademia Nazionale dei Lincei 9:1. Andrea Montella and the Sardinia Autonomous Bartsiokas A, Middleton AP. 1992. Characterization Region for financial support of his Master and and dating of recent and fossil bone by X-ray diffraction. Journal of Archaeological Science 19: 63–72. Back project TS160. The authors thank Prof. Benichou Safar H. 1982. Les tombes puniques de Maria Luisa Ganadu, Dr Giulio Ibba and Dr Carthage. Topographie, structures, inscriptions et Sergio Scognamillo (University of Sassari) for rites fune´raires. Etudes d’Antiquite´s Africaines 237–248. making available the use of FT-IR spectroscopy, Bohnert M, Rost T, Pollak S. 1998. The degree of and Prof. Vittorio Mazzarello (University of Sas- destruction of human bodies in relation to the sari) for supplying the osseous materials duration of the fire. Forensic Science International 95: employed in the FT-IR calibration. 11–21.

Bonucci E, Graziani G. 1975. Comparative thermo- Perinet G. 1964. Determination par diffraction X de la gravimetric X-ray diffraction and electron micro- temperature de cuisson d’un ossement calcine`. scope investigations of burnt bones from recent, Application au materiel prehistorique. Comptes ancient and prehistoric age. Atti della accademia Rendus d’Acudemie des Sciences, Paris (Series D) 258: Nazionale dei Lincei. Rendiconti, classe di scienze fisiche, 4115–4116. matematiche e naturali 59: 517–532. Person A, Bocherens H, Salie`ge JF, Paris F, Zeitoun V, Botto M, Salvadei L. 2005. Indagini alla necropoli Ge´rard M. 1995. Early diagenetic evolution of bone arcaica di Monte Sirai, relazione preliminare sulla phosphate: an X-ray diffractometry analysis. Journal campagna di scavi del 2002. Rivista Studi Fenici 33: of Archaeological Science 22: 211–221. 81–167. Piga G, Malgosa A, Mazzarello V, Bandiera P, Melis P, Enzo S, Parrish W. 1983. A method of background Enzo S. 2007a. Anthropological and physico- subtraction for the analysis of broadened profiles. chemical investigation on the burnt remains of Advances in X-Ray Analysis 27: 37–41. Tomb IX in the ‘‘Sa Figu’’ hypogeal necropolis Enzo S, Bazzoni M, Mazzarello V, Piga G, Bandiera P, (Sassari, Italy), Early Bronze Age. International Journal Melis P. 2007. A study by thermal treatment and X- of Osteoarcheology 18: 167–177. ray powder diffraction on burnt fragmented bones Piga G, Malgosa A, Enzo S. 2007b. Estudio de cre- from Tombs II, IV and IX belonging to the hypogeic maciones espanolas y sardas a trave`s del ana`lisis de necropolis of ‘‘Sa Figu’’ near Ittiri-SS (Sardinia-Italy). difraccio`n de rayos X (XRD). Revista Espanola de Journal of Archaeological Science 34: 1731–1737. Antropologia Fisica 27: 93–102. Gifford D. 1981. Taphonomy and paleoecology: a Piga G, Malgosa A, Thompson TJU, Enzo S. 2008. critical review of archaeology’s sister disciplines. In A new calibration of the XRD technique for the Advances in Archaeological Method and Theory, Vol. 4, study of archaeological burnt remains. Journal of Schiffer M (ed.). Academic Press: New York; 365– Archaeological Science 35: 2171–2178. 438. Rietveld HM. 1967. Line profiles of neutron powder- Guirguis M. 2006. Nuovi dati dalla necropoli fenicia e diffraction peaks for structure refinement. Acta Crys- punica di Monte Sirai (Sardegna). La Tomba 248. tallographica 22: 151–152. L’Africa Romana 17: 1639–1658. Rink WJ, Schwarcz HP. 1995. Tests for diagenesis in Hare PE. 1980. Organic geochemistry of bone and its tooth enamel: ESR dating signals and carbonate relation to the survival of bone in the natural contents. Journal of Archaeological Science 22: 251–255. environment. In Fossils in the Making. University of Rodero Riaza A. 2001. El ritual funerario en las Chicago Press: Chicago; 208–219. necro´polis coloniales andaluzas. In Arqueologıá Hedges R, Millard A. 1995. Bones and groundwater: funeraria: las necro´polis de incineracioń, Garcıá Huerta towards the modelling of diagenetic processes. R, Morales Herva´s FJ (eds). Ediciones de la Uni- Journal of Archaeological Science 22: 155–164. versidad de Castilla-La Mancha, Universidad de Hedges REM, Millard AR, Pike AWG. 1995. Measure- Castilla-La Mancha. 79–90. ments and relationships of diagenetic alteration of Shahack-Gross R, Bar-Yosef O, Weiner S. 1997. bone from three archaeological sites. Journal of Blackcoloured bones in Hayonim Cave, Israel: dif- Archaeological Science 22: 201–209. ferentiating between burning and oxide staining. Knight B. 1996. Burns and scalds. In Forensic Pathology, Journal of Archaeological Science 24: 439–446. (2nd edn), Knight B (ed.). Edward Arnold: London; Shipman P, Foster G, Schoeninger M. 1984. Burnt 305–317. bones and teeth: an experimental study of color, Lee-Thorp J, Van der Merwe NJ. 1991. Aspects of the morphology, crystal structure and shrinkage. Journal chemistry of modern and fossil biological apatites. of Archaeological Science 11: 307–325. Journal of Archaeological Science 18: 343–354. Sillen A. 1989. Diagenesis of the inorganic phase of Lutterotti L, Ceccato R, Dal Maschio R, Pagani E. cortical bone. In The Chemistry of Prehistoric Human 1998. Quantitative analysis of silicate glass in Bone, Price TD (ed.). Cambridge University Press: ceramic materials by the Rietveld method. Materials Cambridge; 221–229. Science Forum 278–281: 87–92. Sillen A, Le Gros R. 1991. Solubility profiles of Michel V, Ildefonse Ph, Morin G. 1996. Assessment of synthetic apatites and of modern and fossil bones. archaeological bone and dentine preservation from Journal of Archaeological Science 18: 385–397. Lazaret Cave (Middle Pleistocene) in France. Palaeo- Sillen A, Sealy JC. 1995. Diagenesis of strontium in geography, Palaeoclimatology, Palaeoecology 126: 109– fossil bone: a reconsideration of Nelson et al. (1986). 119. Journal of Archaeological Science 22: 313–320.

Copyright # 2008 John Wiley & Sons, Ltd. Int. J. Osteoarchaeol. 20: 144–157 (2010) Funerary Rites at a Phoenician-Punic Necropolis 157

Sillen A, Parkington J. 1996. Diagenesis of bones from Selargius-Cagliari. Edizioni La Torre: Cagliari; Eland’s Bay Cave. Journal of Archaeological Science 23: 255–277. 535–542. Wagner CNJ. 1966. Chapter 7. In Local Atomic Arrange- Stiner MC, Kuhn SL, Weiner S, Bar-Yosef O. 1995. ments Studied by X-ray Diffraction, Vol. 36, Cohen JB, Differential burning, recrystallization, and fragmen- Hilliard JE (eds). Gordon and Breach: New tation of archaeological bone. Journal of Archaeologi- York. cal Science 22: 223–237. Weiner S, Bar-Yosef O. 1990. States of preservation Surovell TA, Stiner MC. 2001. Standardizing infra-red of bones from prehistoric sites in the Near East: measures of bone mineral crystallinity: an exper- a survey. Journal of Archaeological Science 17: 187– imental approach. Journal of Archaeological Science 28: 196. 633–642. Wopenka B, Pasteris JD. 2005. A mineralogical Tuross N, Behrensmeyer AK, Eanes ED. 1989. Stron- perspective on the apatite in bone. Materials Science tium increases and crystallinity changes in tapho- and Engineering C 25: 131–143. nomic and archaeological bone. Journal of Wright LE, Schwarcz HP. 1996. Infrared and isotopic Archaeological Science 16: 661–672. evidence for diagenesis of bone apatite at Dos Pilas, Ugas O, Lucia G. 1987. Primi scavi nel sepo- Guatemala: paleodietary implications. Journal of lcreto nuragico di Antas, Atti del II convegno Archaeological Science 23: 933–944.