Stone-cist grave at Kasekula, western Estonia, in the light of AMS dates of the human bones/Kasekula kivikirstkalme Laane-Eestis AMS-dateeringute valguses.

Laneman, Margot

Introduction

The main aim of this article is to present and discuss new radiocarbon (AMS) dates of the human bones collected from stone-cist grave I at Kasekula, western Estonia. The stone-cist grave was excavated by Mati Mandel in 1973 with the purpose of specifying the settlement history of the region (Mandel 1975). So far it has remained the only excavated stone-cist grave in mainland western Estonia (Mandel 2003, fig. 20). Excavation also uncovered a Late Neolithic settlement site beneath the grave, which was further investigated by Aivar Kriiska in 1997 (Kriiska et al. 1997). His 22 [m.sup.2] excavation on the southern and south-eastern side of the grave also exposed the ruin stones of the grave. Based on artefact finds, the construction of the grave is currently dated to the IV or V period of the (Nordic) Bronze Age and the presence of Pre-Roman Iron Age burials has been considered likely (Lang 1996, 297). The human bones of the grave were examined by Jonathan Kalman (2000a) and Raili Allmae (2010), which offers a rare opportunity to compare two expert views on the same osteological data.

Radiocarbon dating of the grave's human osteological material was part of a more extensive project designed to investigate the chronology and burial practices of stone-cist graves, in particular the well-known issue of protracted use of stone graves. Before 2009 there were only four radiocarbon dates of human skeletons from three graves and charcoal dates were hardly more numerous (Lang 2007, fig. 97), although the number of excavated stone-cist graves in Estonia is well over a hundred and dating them on the basis of grave goods is notoriously difficult. Without a good chronological framework, however, palaeodemographic and social inferences would hardly make any sense. To improve the situation, bone samples for AMS dating were selected from several stone-cist graves from which the human osteological assemblage had undergone sex and age determination and which contained burials both in the cists (assumed to date from the Late Bronze or Early Pre-Roman Iron Age) and outside cists (assumed to date from later periods). The grave at Kasekula meets these criteria perfectly.

There already exist three AMS dates for this grave, published by Raili Allmae in 2010. It is necessary to note that her samples were collected before mine, and I was unaware of her research; otherwise my selection of samples would probably have been slightly different. The new dates, however, significantly enhance the existent chronology of the grave and also shed some light on the discrepancy between the osteological studies by Kalman (2000a) and Allmae (2010). To provide a comprehensive discussion, I publish the dates for this grave in a separate article, while the dates for other stone-cist graves involved in the aforementioned project will follow shortly.

In what follows, the description of the grave itself, i.e. its structure, find assemblage, burials, as well as its neighbouring sites come first. This is essential for the interpretation of radiocarbon dates; also, the find assemblage, particularly pottery, has undergone no thorough examination previously and thus deserves extra attention. After that the radiocarbon dates are presented and discussed, including their correlation with artefact finds and osteological estimations, as well as their representativeness and relevance in the wider context of prehistory in Estonia.

Context of the grave

The stone-cist grave under review is located three kilometres from the western coast of mainland Estonia, 600 m south-west of Kasekula village (Fig. 1). Here a north-south aligned elongated ridge, a former coastline formation that is today called Parnamagi [Linden Hill], rises 1-1.5 m above the surrounding fields and pastures. The ridge is densely topped with a row of stone graves, probably stone-cist graves. Altogether seven graves have been registered, but the actual number of graves is larger (see grave descriptions in the National Registry of Cultural Monuments). The grave discussed in this article (grave I) is the northernmost on the elevation, and the only excavated grave of the group. Beneath the grave(s) lies the cultural layer of a Late Neolithic (Late Comb Ware) settlement site, the extent of which is unknown.

[FIGURE 1 OMITTED]

East and south-east of this grave group are at least nine other stone graves scattered between the fields. Many of them are possibly stone-cist graves, a few are tarand graves, and one (grave XIV) is an extensive low stone grave-field with cremation burials, roughly 100 x 40 m in area. Small-scale excavations at this site in 1973 and 2001 yielded cremated bones and a few artefacts from the 11th--12th centuries AD (Mandel 2003, 108). Originally there may have been graves also west and north-west of the village (Mandel 1973, 2).

Other archaeological sites at Kasekula include a 55 cm tall boulder with 39 cup-marks not far from the 11th-12th-century grave-field. Excavation around the cup-marked stone in 1973 yielded a burned enamel bead and a bronze spiral, which appear to associate with the nearby grave, while the original context of cup-marked stones is the Bronze Age and Pre-Roman Iron Age (Lougas 1975). The settlement site north of the graves, at the heart of the present-day village, is probably as old as the stone-cist graves. It has also yielded finds from later periods, which has encouraged opinions that the site may have been continuously inhabited from the Late Bronze Age until today (Lougas 1975; Mandel 2011, 110 f.). 200 m west of the grave group at Parnamagi once lay ten hectares of 8th-12th-century cairn fields, today destroyed by dolomite extraction (Lang 2000a).

Structure of the grave

The grave had been bounded by a circle of large granite stones, the measurements of which reached 50-60 cm (Figs 2-3). The circle, not perfectly ring-shaped, was 12.1-12.8 m in diameter. There was no evidence that the circle had been topped with additional stone layer to form a higher stone wall. The walls of the cist within the stone circle had been built of boulders as large as the boulders of the circle. The floor of the cist had been made of limestone slabs and the roof probably of granite stones. One of the latter had been preserved at the southern end of the cist and was visible even before excavation, suggesting that the cist had been easily accessible. The inner measurements of the cist were 2.2 x 0.5 m, the depth approximately 0.5 m. The space between the cist and the circle was filled with slightly smaller granite stones and limestone, the latter being less frequent than granites. The overall impression was that the stones were slightly larger in the western and southern parts of the grave than in the eastern and northern parts, which can perhaps be explained by the fact that the northern and eastern parts had recurrently been used for burial, while the western part of the grave had not. The thickness of the stone pile of the grave averaged 55 cm (Mandel 1973; 1975).

This quite a typical stone-cist grave adjoined to a rather puzzling structure at its northern side. Three metres north of the stone circle were three big granite stones that appeared to form a southern end of a limestone-floored stone cist, while the middle and northern part of the probable cist had been completely destroyed. The space between the cist remains and the stone circle to the south was full of stones, slightly smaller in size and with a greater proportion of limestone than within the stone circle. Photographs taken during excavation suggest that the granite stones clustered mostly around the cist end while closer to the stone circle limestone became more abundant. This was possibly due to the inhumation burial found 0.9 m north of the stone circle (for detailed information, see the following subsections).

[FIGURE 2 OMITTED]

It is not clear how to interpret this part of the excavated area. Its ambiguous character has, for instance, instigated the suggestion that it was a structure for conducting rituals with human bones, particularly skulls (Jonuks 2009, 176). The excavator of the grave simply refers to an 'extension' [of the grave] (Mandel 1973; 1975) but he does not exclude the possibility that it had been a separate stone-cist grave the majority of which had at some point been destroyed (Mati Mandel, pers. comm. 23.03.2012). Indeed, there would have been enough space for a common stone-cist grave up to 8-9 m in diameter. The fact that no remains of a stone circle or ring wall were found can be explained by the presence of the abovementioned inhumation, which may have destroyed the stone circle in this part of the grave. After all, stone-cist graves with no surrounding circle are possible, although such graves are very rare in Estonia (Lang et al. 2001, 41).

[FIGURE 3 OMITTED]

Considering the possibility that the structure in question was not a separate grave, one would expect it to have been clearly defined with a distinct stone border, in the manner of either a rectangular tarand (e.g. Lang 2007, fig. 103), or, less likely, a stone-cist grave that shares part of its ring wall with another stonecist grave (e.g. Lang 2007, fig. 99). These possibilities are not entirely excluded at Kasekula, although in that case the southernmost sections of the expected stone border must have been observable in the excavation area--but they were not. Ambiguous stone structures that are attached to a stone-cist grave, contain cists, and have no clearly defined stone surround have so far not been recorded in Estonia, although this fact does not exclude the possibility of their existence.

Although the following cannot offer a definitive statement on the character of the stone structure under review, one might nevertheless notice that some of the neighbouring graves at Parnamagi seem to be accompanied by similar structures (see the descriptions in the National Registry of Cultural Monuments). For instance, grave IV (register no. 9838) has been described as featuring a low rectangular hump ca 3 x 4 m in size attached to its southern side, which encourages the inference of 'extending' the graves. On the other hand, the rather large grave III (9837) adjoins a smaller circular mound, 5.6 m in diameter, and a similar small hump was located 2 m south of grave I. The latter have been assumed to be separate graves (although not registered as such), suggesting a possible pattern of smaller graves located between larger ones.

Another peculiar feature worth mentioning was recorded on the south-eastern side of the grave (Fig. 3; Kriiska et al. 1997, 30 f., 38). It was a 25 cm deep oval pit dug into the subsoil, 140 x 85 cm in size. The pit was filled with soil and stones, which showed clear indications of burning: granites were fire-cracked and limestone had turned red or even calcified. The hollow contained a few pieces of animal bones and pottery, at least some of the latter most probably of a Bronze Age date (see below). A similar feature with fired stones and charcoal, over 30 cm in depth and 80 cm in diameter (tapering downwards), has been recorded on the southern side of one of the stone-cist graves at Karuste, Saaremaa (Vassar 1941; cf. Lang 2000b, 99). Vassar (1943, 17) compares the structure with similar features beneath tarand graves and proposes their association with rituals performed during the grave construction.

Artefact finds and animal bones

Characteristically of stone graves, the finds were scattered between the stones and it was not possible to associate them with particular burials (Fig. 3). Here I discuss only the finds that are later in date than the Neolithic; for the find assemblage of the Stone Age settlement site, refer to Kriiska et al. 1997. The finds are stored in Estonian History Museum under registry numbers 476 and 807.

The only artefact find in the central cist was a bronze razor with a broad handle, concave back and a small rounded protrusion on the blade side (Fig. 4: 2). Such items probably originate from Jutland, Denmark, and date from period IV or the beginning of period V of the Nordic Bronze Age (Baudou 1960, 36 f.). The item most similar to it found in Estonia was at the cist of grave 19 at Joelahtme, Harjumaa, along with the remains of an adult male (Kraut 1985, pl. V: 8). The bones of the male have been radiocarbon-dated to approximately 930-800 BC (unpublished data in the possession of author). Stone-cist graves in Estonia have yielded altogether eight bronze razors or small knives: besides Kasekula and Joelahtme 19, there are an additional three from Joelahtme, one from Kangru IV at Vao, Harjumaa, one from Sepa at Kaarma, Saaremaa, and one from Karuste (Suurevare), Saaremaa. Three of the respective graves at Joelahtme were burial places for adult men and one of them also contained scant remains a 10-12-yearold child; the fourth grave, however, contained an 8-9-year-old child and an infant (Kalman undated). The bones of all three men were AMS-dated; the result for one of them was mentioned above, while the calibrated (95.4%) dates for the remaining two range from 1260-940 BC (unpublished data of author). Although commonly referred to as 'razors', the function of the items is disputable and may have differed by region (see Thedeen 2003).

[FIGURE 4 OMITTED]

The head of an iron shepherd's crook pin (Fig. 4: 3) was unearthed ca 1.3 m south-west of the central cist at a depth of ca 30 cm; the shaft of the pin was broken and was absent. The pin had been made of an iron rod with a rectangular cross-section, while the majority of such pins appear to have circular or lozenge cross-sections. The most common context for iron shepherd's crook pins are early tarand graves, which sometimes contain dozens of such pins. The peak of their occurrence falls within the Late Pre-Roman Iron Age (200 BC--AD 50), although some may have a slightly earlier date and most certainly this pin type continued in use through the subsequent centuries (Lang 1996, 288 f., 301 ff.; 2007, 183 f.; Laul 2001, 126). The Pre-Roman Iron Age shepherd's crook pins were worn by both men and women, both singly and in pairs (Lang 2007, 183 f.); no child burials with such pins can be mentioned, but this may be due to shortage of osteologically examined bone assemblages. It is thought that the pins were used for fastening clothes, probably at the shoulders (Schmiedehelm 1952, 17; Lougas 1984a, 349, pl. I: 2); later pins also functioned as chain holders (Laul 2001, fig. 50). The frequency of the iron shepherd's crook pins indicates their local production.

Among the excavated stone-cist graves of Estonia (including disputable cases of circular graves with no (i.e. destroyed?) cists), around twenty contain shepherd's crook pins; at least five of them yielded a bronze pin. The former figure makes up less than 15% of the above-mentioned grave category. A few of the pins are younger than the Pre-Roman Iron Age, while the majority probably fall within the Late Pre-Roman Iron Age. Making more precise estimations is problematic, however, and would require the compilation of a new special study of this pin type (Lougas 1971 is outdated). It is, however, clear that shepherd's crook pins in stone-cist graves represent burials (or offerings?) interred many centuries after the construction of the graves. It is common that they occur outside cists and cannot be associated with a particular burial (with only one firm exception, and even then it was not the original cist of the grave, see Lougas 1984a). In many cases there is only one shepherd's crook pin in a stone-cist grave, but graves with a greater number of pins also occur.

At 1973 excavation 110 (1) pottery sherds were collected (Mandel 1973), at least 80 of them with a Metal Age date. The great majority of the sherds were found singly or paired; only two larger clusters were observed, one in the northwestern periphery of the grave (30 sherds) and the other in the southern periphery of the grave (10 sherds). The sherds are mostly small fragments from the vessels' walls, while rim fragments are rare and bottom fragments are altogether absent, which makes the reconstruction of shape and dating of the vessels very complicated.The 1997 excavation south of the grave yielded 83 Metal Age potsherds from a minimum of five vessels (Kriiska et al. 1997, 31 ff.), but some of them may be associated with the grave less than 2 m south of grave I. The pottery assemblage was examined by Prof. Valter Lang (except for the sherds collected in 1997, which were browsed in a cursory manner).

Within the stone circle, four vessels could be distinguished relatively well. One of them had been next to (originally on top of?) the stone circle in the northwestern margin of the grave where a cluster of 30 potsherds was found, with two more sherds of the same vessel slightly apart, one of them outside the stone circle (squares 6-7/m, see Fig. 3). It was a coarse ware clay pot with somewhat everted rim and a decoration of elongated, slightly oblique grooves on its shoulder (Fig. 4: 4). A similar, although not identical vessel is shown in Lang 1996, fig. 9: 8. Such vessels have usually been dated to the Pre-Roman Iron Age, but a Late Bronze Age date cannot be excluded (Lang 1996, 43). Another coarse ware clay vessel was found a few metres south-east, but it was represented with only five or six wall fragments, which were scattered 1-2 m apart from one another (squares 7/n and 8/m-n) and do not enable the shape of the vessel to be reconstructed. The latter applies also to the third vessel, which was located in the southernmost periphery of the stone circle (11/o) and was represented by roughly a dozen coarse-grained sherds. A Pre-Roman Iron Age date is likely for this vessel. The fourth vessel was located a few metres north-east of the latter (11/p, perhaps also 10/p-q) and was represented by less than ten scattered sherds. This vessel had once had a smoothed surface and concave neck, but the shape of the rim could not be determined. The date of the vessel within the Late Bronze Age and Pre-Roman Iron Age cannot be specified.

The rest of the Metal Age pottery within the stone circle, mostly in the northeastern sector, consisted of indefinite sherds often mingled with Stone Age sherds. Worthy of separate mention is a tiny single rim fragment at the western side of the cist that shows cord impressions on its lip. Vessels with various imprints on the flat parts of their rims are relatively numerous in the Pre-Roman Iron Age contexts (Lang 1996, 143).

Only one sherd out of six found north of the stone circle deserves more attention (Fig. 4: 5). It was found near the inhumation burial in square 5/p, although from a slightly higher context. The relatively large sherd originated from a coarse ware pot with a smoothed surface, curved walls 5.5 mm in thickness and a slightly everted flat-topped rim (similar in terms of shape to a pot in Lang 1996, fig. 12: 5). This pottery type is common in early tarand graves, and Lang confidently dated the sherd in question to the Late Pre-Roman Iron Age through the Roman Iron Age.

The pottery found south of the grave is apparently similar to the pottery within the stone circle, representing coarse-grained vessels with thick walls and often with smoothed surfaces (Kriiska et al. 1997, 31 f.). Two fragments, however, deserve more attention, because they are more clearly attributable to the Bronze Age than any other potsherd in the excavated area. The sherds are suggestive of a relatively large vessel with a smoothed surface and striated interior; the rim was rather straight (Fig. 4: 6) and the shoulder probably smoothly curved. Such vessels have been classified as the 'Asva-style coarse-grained ceramics' and can be dated to the end of the Early Bronze Age through the Late Bronze Age (Lang 2007, 127 f.). The sherds were found in the pit at the south-eastern side of the stone circle.

To sum up, the excavated area contained pottery from both the Late Bronze Age and Pre-Roman Iron Age, but the respective proportions are fairly difficult to establish. A look at other stone-cist graves reveals that pottery is not an unusual find category in them. The cists sometimes contain the so-called Luganuse-style pots, which have been dated to the Late Bronze and Pre-Roman Iron Ages, although the few available radiocarbon dates favour the earlier part of the time range (Lang 2007, 129 f.). Outside cists, the pottery is considerably less well-preserved and a pattern similar to Kasekula is not unusual. It nevertheless seems that the pottery found outside cists in many cases belongs to the same Luganuse style (see Lang 1996; in the latter study this pottery is classified as type BII: b) and in a few cases represents Asva-style coarse-grained (BIII) or Asva-style fine-grained ceramics (AI). The latter styles have been dated by Lang to the Bronze Age (rather than Pre-Roman Iron Age) with a greater confidence than the Luganuse-style pottery (Lang 1996; 2007, 127 ff.).

One might conclude from this that pottery outside cists may in many cases be of a Bronze Age date. Despite this, usually a Pre-Roman Iron Age date is preferred, because finds and burials outside cists are believed to be of a younger date than finds and burials inside cists (Lang 1996; 2007, 160). The latter assumption is generally true, and one cannot deny that a few graves do contain pottery that clearly dates from the Late Pre-Roman Iron Age or the later periods, yet the case with pottery may be more complicated than previously thought. First, dating finds on the basis of their location in the grave is an insecure method and second, the connection between the (later) burials and pottery is not entirely clear. In many graves (even at Kasekula) the locations of potsherds and bones do not overlap; there are graves where pottery is found in the very margins, which means that the vessels may have had no direct connection with funerary rites. Further, there are graves that contain pottery but no burials outside the cist, and there are graves where the situation is reversed, which again indicates that a clear-cut relation between pottery and burials does not exist. There is also an opinion that pottery outside the cists is mainly associated with cremation burials (e.g. Lang 2007, 160), a view challenged by the evidence from Kasekula. All things considered, we are left with the conclusion that the dating of the pottery outside the cists of stone-cist graves is based on insecure grounds, and ascribing the fragmentary pottery at Kasekula to a firm Pre-Roman Iron Age group should rather be avoided.

Excavation yielded three stones that may have been used as grinding tools (Fig. 4: 7--8). At least two (2) of them were found at a considerable depth, 55 cm or even deeper (Mandel 1973), which means that they either belong to the Stone Age context or were put in their place at the time of grave construction at the latest. A Stone Age origin, however, seems unlikely (Aivar Kriiska, pers. comm. 17.04.2012), although such stones may have had various functions apart from grinding grain (Vassar 1938, 361; 1943, 234 f.; Lidstrom Holmberg 1998). They have also been found in several other stone graves, for instance a Late Bronze Age ship grave at Lulle, Saaremaa (Lougas 1970) and several Early Iron Age tarand graves (Vassar 1943, 231; Lang 2000b, 123; 2007, 109; Laul 2001, 163), which shows that they are not extraordinary for a grave context. I nevertheless know of only two stone-cist graves beside Kasekula that have yielded grinding stones: grave 6 at Lagedi, Harjumaa, where a grinding stone was found together with Roman Iron Age artefacts (Spreckelsen 1927, 19; Lang 1996, 217), and grave 70 at Muuksi, Harjumaa, where a few grinding stones were found in the ruin of the grave, slightly outside the stone circle (Vassar 1938; cf. Vedru 1998, 52). Here, however, one must consider that grinding tools (as well as other stone items) in stone graves tend to be rarely published, let alone discussed, as the case of Kasekula also demonstrates. These circumstances hardly allow anything beyond a very broad generalisation, as presented in Lang 2007, 109 f. A special study on grinding stones in graves, including use-wear analysis, may produce interesting results (cf. Lidstrom Holmberg 1998).

A small handful of iron slags, altogether six pieces, was found near the skeleton north of the stone circle, at a depth of 35 cm. This is another unspectacular find category that is easy to neglect both at excavation and in publication of excavation results, which makes it difficult to relate such finds to the prehistoric grave context. Artur Vassar, however, argued that iron slag in graves functioned as a grave good or offering (Vassar 1943, 230). The slags found in the Roman Iron Age tarand grave at Nurmsi, central Estonia, originated from iron working (forging), but the presence in graves of slags as a result of iron smelting cannot be excluded (ibid.). As for stone-cist graves, iron slag has been reported for instance from Pihla II at Kopu, Hiiumaa, where it was found above the northern end of the central cist (Lougas 1984b).

The grave yielded a few other ambiguous finds with unknown date and function. Two small fragments of an iron object were found near the eastern side of the central cist at a depth of 22 cm, and the eastern part of the grave yielded a tiny thin bronze disc with a hole in its centre. The latter was found at a depth of 40 cm from the grave surface.

Finally, a short note should be made on animal bones, although the bones from the grave and from the much older settlement site are difficult to differentiate. It could nevertheless be established that the grave material was clearly dominated by cattle, followed by goat or sheep, while dog, pig and horse bones were less numerous (Mandel 1973, 6; Kriiska et al. 1997, 37 ff.). Part of the bones of terrestrial wild animals and various fish species found in the stone ruin south of the grave were also primarily associated with the grave rather than the settlement site, although the relatively numerous seal bones were mainly attributed to the settlement site (Kriiska et al. 1997, 38).3 A largely similar pattern of animal bones, including seal bones, is known from Late Bronze Age fortified settlements, Late Bronze-Early Iron Age stone graves and also from coastal Estonian stone graves from later periods of the Iron Age (Vassar 1938, 332; Lang 2000b; 2007, 110 f.; Maldre 2000; Tvauri 2012, 106, 285 f.). Interestingly, the central cist has been said to have contained bones from a paw of a large animal, which was interpreted as an animal hide placed over the deceased (Mandel 2011, 112). The distribution of other animal bones in the grave has not been published and nor is it indicated in the excavation report of 1973.

Burials

The human bones found in the grave were unburned. Characteristic of stone graves, the bones were highly fragmented and scattered (Fig. 3). At the excavations in 1973, archaeologists were able to discern a disturbed inhumation in the cist, probably lying with its head to the north, and a poorly preserved skeleton 0.9 m north of the stone circle. The latter was observed as lying on a limestone paving at a depth of 45 cm from the grave surface, with its head to the east and hands stretched over the head. A few human bones were recorded in the destroyed cist north of the stone circle and as scattered elsewhere in the grave, particularly in the vicinity of the central cist, but in these instances archaeologists were able to discern neither buried individuals nor clear-cut burial structures (Mandel 1973, 4 f.; 1975, 74).

Close osteological study of the human bones by Jonathan Kalman (2000a, 18 ff.) about quarter of a century after the excavation revealed that the grave had contained the remains of at least 23 individuals (Table 1). The central cist enclosed a male, at least 50 years old at death and approximately 178 cm tall, and a minimum of three infants less than one year of age. Reportedly, the man had suffered from osteoarthritis and severe osteophyte formation on thoracic vertebrae, which would have caused limping when walking and stiffness in the back. Among what were recorded as scattered bones outside the central cist within the grave circle, Kalman was able to discern the remains of a 3-4-year-old child and at least 16 infants in their first year of age, with considerable probability that the original number of infants was even greater. North of the grave circle, Kalman detected at least two adults. One of them, represented by only fragments of fibula, distal right femur and radius, had apparently been lying in the destroyed cist; its age and sex were indeterminable. The skeleton found between the destroyed cist and the stone circle turned out to be a male over 50 years in age; the bones were fragmented but almost all parts of the body were present.

In her recent re-examination of the bone assemblage, Raili Allmae (2010) concluded from the number of petrous parts of the temporal bones that the excavated area contained at least 28 individuals (Table 1). She agreed with Kalman with regard to the man in the central cist, but found four infants instead of three in this cist. The number of infants outside the cist by Allmae is five more than Kalman's estimate. Her measurements of infant long bones show that the buried infants died within at most six months after birth, which is a more exact estimation than Kalman's twelve months. Allmae also confirmed the presence of a child within the stone circle and a man older than 50 years outside the stone circle, but rejected the presence of the second adult in the latter grave area, arguing that this individual was distinguished by Kalman on the basis of only one piece of cranium (Allmae 2010, 47).

This reading of Kalman is, however, incorrect, as he clearly lists the present parts of the incomplete skeleton found in what was probably a destroyed stone cist (Kalman 2000a, 20; see also above). Also, a look at the excavation plan (Fig. 3) suggests that at least two different inhumations can be expected from the northernmost part of the excavated area, one from the destroyed cist and another south of it. Allmae's negligence about these circumstances is perhaps due to her focus on the infant burials.

Allmae collected three samples for radiocarbon dating in Poznan Radiocarbon Laboratory; the sample information and results are presented in Table 2.

New AMS dates

In September 2009 when collecting the samples for radiocarbon dating, we proceeded from the skeletal study of Kalman (2000a), being unaware of Allmae's research. Bioarchaeologist Martin Malve from the University of Tartu matched the bones in storing boxes with individuals distinguished by Kalman; a systematic revision of the bone assemblage was not undertaken.

The main question of interest with this grave was: what is the temporal distance between (1) the burials in the central cist, (2) burials outside the cist, and (3) burials outside the grave circle? Should they be considered in the same or different temporal contexts? Proceeding from this question, we collected nine samples for radiocarbon dating from different parts of the grave; the samples most probably represent eight individuals. The sample information along with the results of AMS dating is summarised in Table 3 (see also Figs 3 and 5). The samples were processed by the dating laboratory of the Finnish Museum of Natural History; along with [sup.14]C analysis, carbon and nitrogen stable isotope ratios were also measured (though thorough discussion of the latter results remains beyond the scope of this article).

The male in the central cist is represented with two samples (Nos 1 and 2 in Table 3); (4) when including Allmae's data, this number increases to three. Unsurprisingly, this inhumation turned out to be one of the oldest in the whole lot. All three radiocarbon estimations by two different laboratories agree that the man had lived before 800 BC; the minor differences between the radiocarbon ages fall within the limits of statistical error. The radiocarbon dates are in accordance with the date of the bronze razor found in the cist.

Two samples are associated with the destroyed cist north of the stone circle. One of them (No. 3 in Table 3) originated from bones found directly from between the three large stones of the cist end and most probably represents the incomplete skeleton of an adult person discussed by Kalman (see above). Radiocarbon dating indicates that this person is more or less contemporaneous with the male in the central cist, i.e. earlier than 800 BC in date. The other sample (No. 4) comes, according to the label that accompanied the bones, 'from the limestone paving next to the remains of the cist', i.e. apparently slightly outside the cist end (from which side was not specified). (5) It was not possible to determine whether or not these bones belonged to the individual inside the cist end. Given that Kalman was rather ambiguous about the number of persons inhumed in this part of the grave, radiocarbon dating was ordered on the bones to see if they represent a different individual. Considerable difference in BP estimations suggests that they do, although in reality this individual need not be much younger in date than the other individual in the destroyed cist.

[FIGURE 5 OMITTED]

The radiocarbon dates from this part of the grave thus support the osteological determinations of Kalman rather than Allmae, since the latter was not able to discern any individuals associated with the destroyed cist (see above). The dates also show that the destroyed cist is not much younger, if younger at all, than the grave south of it, which perhaps supports the inference of a distinct (destroyed) grave rather than a later extension of the southern grave. Radiocarbon dates, however, cannot provide a conclusive solution for the latter problem.

The male skeleton north of the stone circle (between two graves?) recognised by both Kalman and Allmae is represented by one sample (No. 5). Radiocarbon dating suggests that the man had been interred at least one hundred years after construction of the grave, although a much later date close to the boundary between the Bronze and Iron Ages appears to be more plausible. According to Valter Lang, the fragment of a clay vessel found near the skeleton is not older than the Late Pre-Roman Iron Age, which clearly disagrees with the radiocarbon estimate. Iron slags, also found in the vicinity of the skeleton, cannot be dated, but they might be viewed as an appropriate grave inclusion in the earliest Iron Age when iron was a rarity. Neither of the finds can assuredly be associated with the burial under review, however. The grave also contained inhumations radiocarbon-dated to the Late Pre-Roman Iron Age (see below) and the presence of the potsherd in question is thus not entirely unaccountable.

A fragment for sampling (No. 6) was selected from the bone cluster near the stone circle in the south-western periphery of the grave. The bones had been lying no deeper than 25 cm from the grave surface (Mati Mandel, pers. comm. 07.04.2012). Neither of the osteological studies commented on this bone assemblage, despite its rather separate location in the grave. According to Martin Malve however, it was an extremely incomplete inhumation of an adult. (6) Radiocarbon dates the bones later than 200 BC, which confirms that they cannot be attributed to any other adult burial identified in the grave.

Three samples were collected from the bones of subadults outside the central cist. (7) One of them represents the 3-4-year-old child (No. 7) (8) and the remaining two represent infants (Nos 8-9). The bones of the infant near the central cist had been lying close to the grave surface, at a depth of less than 25 cm, while the remaining two youths had been unearthed from slightly deeper layers, more than 30 cm from the grave surface (Mati Mandel, pers. comm. 07.04.2012). Radiocarbon suggests that the child in the eastern part of the grave is roughly contemporaneous with the adult in the south-western periphery of the grave, i.e. younger than 200 BC, while the infant in the south-eastern part of the grave is of a slightly earlier date, between 400 and 200 BC. The infant near the cist corner, however, dates from the Viking Age and belongs with the same group of two Late Iron Age infant burials radiocarbon-dated by Allmae.

The Late Iron Age use of the grave had not been indicated by any artefact find, whereas Pre-Roman Iron Age burials were expected due to the presence of an iron shepherd's crook pin and at least some of the pottery. These finds generally agree with the radiocarbon dates, which span almost the entire Pre-Roman Iron Age. Narrowing down this date range or specification of date of the artefact finds is difficult, partly because the relation of the finds to burials is ambiguous. The shepherd's crook pin probably dates from the second half of the Pre-Roman Iron Age, although a slightly earlier date cannot be precluded. Dating of the pottery within the Late Bronze and Pre-Roman Iron Ages was shown above to be extremely difficult. Although the radiocarbon dates suggest that at least part of the ambiguous potsherds found within the stone circle may now be dated to the Pre-Roman Iron Age rather than the Bronze Age, one should still be cautious with extending the date to the whole pottery assemblage.

Burial sequence at Kasekula

The results of radiocarbon dating show that both cists originate from the first half of the Late Bronze Age, while inhumations outside the cists are divided between the Pre-Roman Iron Age and Late Iron Age. Such a temporal distribution of the deceased clearly suggests discontinuity in burial. It is also obvious from this pattern that the social position of a deceased cannot be inferred from its location in the grave (cf. Jaanits et al. 1982, 200), because burials of separate periods must be addressed independently in the contexts of respective periods. However, in focusing on these periods one must consider that we cannot determine how those 15-21 infants that were not radiocarbon-dated distribute between the periods. Also, a group of stone-cist graves as a creation of a single community (i.e. a family cemetery or shrine) should be viewed as a whole, which means that many important questions cannot be answered unless at least the majority of the group has been excavated.

The most likely date of construction for grave I at Kasekula can probably be located between 950 and 850 BC, while the destroyed cist, perhaps part of a separate grave, cannot be much younger. In view of the general chronology of stone-cist graves in Estonia it is safe to assume that grave I is one of the oldest stone graves at Kasekula, i.e. there cannot be many graves that predate it. Based on the current evidence it seems most likely that burial in the excavated grave(s) stopped before 700 BC. Extension of this inference to the whole grave group at Parnamagi, let alone farther graves is, however, much more questionable and should rather be avoided. One must also consider that some of the infant burials with no radiocarbon date, both in and outside the cist, may belong to the Bronze Age, a probability proved by evidence from other graves (unpublished data of author). Although such circumstances would probably not considerably alter the above chronological estimations, one should nevertheless avoid concluding that in Bronze Age Kasekula only adults were eligible for burial in a stone grave and only cists were used for interment. In any case, evidence from only one grave of the group cannot determine whether all or only selected members of the community were accorded burial in a stone grave.

Burial in the grave began again in the Pre-Roman Iron Age. The radiocarbon dates of these inhumations span the whole period beginning at 500 BC, although the actual period of use may have been shorter, for instance 400-150 BC. The buried individuals include both adults and children of different age categories, and it is notable that the number of burials is larger than just occasional one or two. Precise numbers remain unknown, since without further radiocarbon dating there are no means to establish how many of the infants have a Pre-Roman Iron Age date. In general, there are no obvious reasons to argue that the burials under review considerably deviate from patterns observed in early tarand graves, the most common stone grave type in Pre-Roman Iron Age Laanemaa. These graves also mainly contain inhumations of both adults and children, accompanied with no or a few grave goods (e.g. Lang 1996, 302 f.; 2007, 173 ff.; Kalman 2000a; Mandel 2000).

It is not easy to judge whether or not the Pre-Roman Iron Age burials at Parnamagi represent the kindred descendants of the community that once built the graves and whether the graves served as their only, main or secondary burial ground. Although the available AMS dates suggest a few hundred-year-break in burial before the Pre-Roman Iron Age, it is still possible that burials of this interim are to be found in other graves at Parnamagi. (9) In other words, one cannot entirely exclude the possibility that use of the cemetery and existence of the community were continuous from the Late Bronze through the Pre-Roman Iron Age, although the meaning of the grave structure perhaps altered with the passage of time.

On the other hand, discontinuity in burial may have been real, although this need not mean discontinuity in the existence of the settlement unit-there are many other graves in the neighbourhood that may have served in the interim period. The nearby settlement site was perhaps also continuously inhabited. In the Pre-Roman Iron Age the community possibly buried most of its dead in tarand graves, which are present at Kasekula and most probably belong to the Pre-Roman Iron Age, since Roman Iron Age tarand graves are rare in western Estonia (e.g. Mandel 2000, 107 f.; Lang 2007, 91 f.). Old stone-cist graves were perhaps used for interment of only a few of the community members, although there are no signs to identify what the criteria for such a differentiation were. Scenario like this seems to have been at work at Rebala, where only two out of five graves contained Pre-Roman Iron Age burials and their number in relation to the Bronze Age burials was small (unpublished data of author). This can be viewed as providing additional grounds to assume that the case with Kasekula is similar, although there are still too many indefinite variables to definitively exclude any other possibility.

In the Late Iron Age, almost two thousand years after its construction, the grave was re-used as a burial place primarily for infants. The available three radiocarbon dates span over many hundred years from the Pre-Viking Age until the end of the Final Iron Age, although the actual burial period may have been limited to a century and half at the end of the Viking Age; a much shorter period is unlikely while a longer period plausible. The number of burials cannot be ascertained.

A similar situation was observed at Rebala where the north-western part of grave II, built before 800 BC, contained at least nine infants in their first year and one 8-9-year-old child (Lang et al. 2001). Radiocarbon dating of the bones of four infants locates their short lives in the 15th-17th centuries AD, perhaps exclusively within the 15th century (unpublished data of author). Other graves in this group did not yield such burials and subadults in their cists appear to be of a Bronze Age date. (10) Infant burials much later in date than the grave itself may also have been present in stone-cist grave IX at Lagedi, Harjumaa (Spreckelsen 1927, 24 f.). Other stone-cist graves in Estonia have not been found to have enclosed such a quantity of infant burials. Many of them have not undergone any close osteological study, but there are reasons to believe that the phenomenon of massive infant burials in stone-cist graves is rather exceptional.

It thus seems that we are observing a case of excluding a certain category of people-infants, perhaps newborns-from the common burial ground (or some other place(s) of disposal-see Lang 2011), while an old stone grave served as the alternative place for interment. If we consider cremation grave XIV a few hundred metres away (Fig. 1) as the main burial ground of the community, then even the mode of disposal appears to have been different (i.e. inhumation instead of cremation). The practice of differential burial of infants need not have been uncommon in Late Iron Age Laanemaa, as thoroughly studied cremation graves appear to yield child burials in considerably less numbers than expected (Allmae 2010, 46, 49, and references therein). The case of Rebala suggests that the practice was much more widespread both in space and time, yet the respective boundaries would be difficult to establish. Apart from Kasekula and Rebala, there are no proved cases in Estonia, and the under-representativeness of child burials which, by the way, has been observed from the Pre-Roman Iron Age tarand graves at the latest (Kalman 2000a, 33; 2000b, 389; 2000c, 429) to the mediaeval rural cemeteries (Valk 2001, 65), does not necessarily imply separate burial places for subadults. Also, infants have rarely been considered separately from older children, a circumstance that tends to obscure potential age-based distinctions in mortuary treatment (Finlay 2000, 413, 418 f.; see also Murail et al. 2004). The subject has thus potential for further research which should perhaps begin from reviewing the tarands where subadult bones predominate (e.g. Kalman 2000c, 425; Lang 2000b, 132) and graves such as Iru II, which was an irregular heap of limestone with 11th-century artefacts, one adult burial and a quantity of infant remains (Lougas 1976). Also relevant are sites such as clearance cairn 8 at Iru, which enclosed the remains of an infant and a slightly older child (ibid., 51). It must be considered, however, that old stone monuments need not have been the only places to seclude dead infants (e.g. Finlay 2000, 411).

Cremation of children is known to be arduous due to their great moisture content, but apart from practical consideration there were most probably religious or ideological grounds for their distinctive treatment in death. It is well known that in many cultures all over the world infants that had not gone through certain initiation rites were not considered fully legitimate members of the community and were thus entitled to a different mortuary treatment, often including exclusion from the community cemetery (Jurgenson 1998, 32 f.; Finlay 2000; see also Paulson 1997, 133; Boric & Stefanovic 2004, 542). In many cases it was some longabandoned monument that was used instead, and it is not uncommon that the site was also used for burying other people whose life course deviated from the norm or who were strangers. Estonian oral tradition also holds that in the final stage of their life-cycles non-consecrated rural cemeteries were used to inter only small children, old people, the poor, beggars, unknown wanderers, Jews and Gypsies (Valk 2001, 91).

At Kasekula, it would be impossible to establish which crucial milestone the deceased infants presumably failed to reach. Christian baptism is unlikely, particularly if the connection with the cremation cemetery is correct, but there are other similar rites or events that may have been critical for being accepted into a society (see Jurgenson 1998, 32). It must also be noted that a separate burial need not necessarily imply a careless or insensitive attitude towards infant deaths in the society (Murail et al. 2004; cf. Kalman 2000a, 20 f.) and one should be cautious with ascribing prehistoric infants the status of outcast. An additional reason for avoiding this perhaps lies in the fact that some Early Iron Age stone graves also contain richly provisioned Viking Age adults, both inhumed and cremated, who could hardly be viewed as outsiders (Tvauri 2012, 255 f.). The best example is perhaps a 10th-century male inhumation with weapons and ornaments in the early tarand grave at Iila, Virumaa (ibid.). On the other hand, there is no reason to believe that all old stone graves had a similar meaning or importance in the Late Iron Age ritual landscape.

There are other interpretations to consider (see Kalman 2000a, 21). Hypothesis of an epidemic is unlikely, one of the reasons being that the radiocarbon dates span too long a period and their overlapping is sparse. There is nevertheless a possibility that all the infants died as a result of a certain disease, perhaps feared as particularly dangerous, but this remains to be proved. Human sacrifice is thought to have been practised in prehistoric Estonia (e.g. Eisen 1996, 8 ff.; Lougas 1996, 86 ff.; Jonuks 2002) but the character and extent of the practice are unclear. Prehistoric child sacrifices are not mentioned in chronicles and the cases observed in the archaeological record (Jonuks 2002 and references therein; cf. Jonuks 2009, 178, 317) raise questions as why they should be viewed as sacrifices. Ambiguous reports on child sacrifices are known from historical period and the motifs also occur in folklore (e.g. Eisen 1996, 10 f.; Lougas 1996, 87), but recorded instances of actual child sacrifices are unknown. If it was infanticide based on some other grounds then the practice must have been accepted in the society, otherwise the great number of infanticide victims in one location would be difficult to explain. A wide range of reasons would apply for such practice, including preference for children of a certain sex, birth defects or other features regarded as abnormalities (e.g. newborns with teeth, twins), birth out of wedlock, shortage of food supplies, etc. (Jurgenson 1998, 31 f.). Perhaps further biological studies, for instance with the aid of DNA studies, would reveal evidence to support the hypothesis of infanticide, but until then and in view of the high infant mortality rate of the time preference should perhaps be better given to other interpretations.

As for the mortuary practices at Kasekula between the Early and Late Iron Age, there are enough graves to bridge the gap. Although registered as stone-cist or tarand graves, they may also include cairn graves and stone grave-fields characteristic of the Middle Iron Age; there is also a possibility for invisible underground (cremation) burials. What really happened remains to be discovered or imagined.

Conclusion

As a result of two independent studies, the number of radiocarbon-dated inhumations at Kasekula is currently ten, which makes up roughly a third of all inhumations uncovered so far. For one of the inhumations there are three radiocarbon dates, which prove the method to be credible. Interpretation of the results is limited due to the fact that the great majority of the grave group has not been excavated, but the case no doubt provides solid material for comparison for other analogous studies which, judging from the case of Kasekula, would potentially yield rewarding results.

Three of the sampled inhumations date from the Late Bronze Age, four from the Pre-Roman Iron Age, and three from the Late Iron Age. Two former periods of use had also been indicated by artefact finds which are in good accordance with the AMS dates, while the latter period was only revealed through radiocarbon dating. These were probably separate burial periods in the use-life of at least the excavated grave, but it is not impossible that the statement applies to the whole grave group at Parnamagi. In any case the burials of different periods must be addressed independently, in their respective chronological contexts; in each of the periods the use and, respectively, the meaning of the grave had been different. Reaching the content of the latter is obviously difficult if not impossible, as there still are a lot of indefinite variables. For instance, the available data does not reveal whether or not all of the members of the community were eligible for burial in a stone grave--a central concern in the research of stone-cist graves. For this the length of use and number of burials of the whole cemetery must be known, but at Kasekula we even do not know how many of the infant burials originate from the Pre-Roman Iron Age and how many from the Late Iron Age; determining the proportions of their chronological distribution would require many more radiocarbon dates than are currently available.

Perhaps the greatest contribution of the discussed radiocarbon dates is to the research into the Late Iron (particularly Viking) Age. The use of an almost 2000year-old grave for burying exclusively infants opens up a range of (much exploited) perspectives, such as the Viking world view, social construction of childhood, stepwise childhood, missing prehistoric children, past in the past, activating the monuments, etc. It should, however, be noted that in regard of the quantity of infant burials the grave at Kasekula is rather an exception than a rule, and further studies are expected to reveal more about the Bronze and Early Iron Age use of the graves. It is in any case advisable that scientific dating of the bones becomes a more regular practice in the research of prehistoric graves and, also, that osteologists have patience to pay attention and clearly discuss the observations of archaeologists made during excavation.

Acknowledgements

This study was financed by the European Union through the European Regional Development Fund (Centre of Excellence in Cultural Theory). I am grateful to Prof. Valter Lang for his help, which ranged from finding financing for the project to the examination of the pottery of the grave. Martin Malve's help has been invaluable in collecting the bone samples for radiocarbon dating. The article has also profited from discussions with Mati Mandel and Prof. Aivar Kriiska, the excavators of the site. Thanks are extended to Krista Sarv and Kristel Kulljastinen for their help in gaining access to the find assemblage of the site and preparing the illustrations for the article, respectively. Mara Woods kindly adjusted my English.

doi: 10.3176/arch.2012.2.01

References

Allmae, R. 2010. Some remarks on Kasekula stone-cist grave, Laanemaa, Estonia.-Fennoscandia Archaeologica, XXVII, 45-52.

Baudou, E. 1960. Die regionale und chronologische Einteilung der jungeren Bronzezeit im Nordischen Kreis. (Acta Universitatis Stockholmiensis. Studies in North-European Archaeology, 1.) Almqvist & Wiksell, Stockholm, Goteborg, Uppsala.

Boric, D. & Stefanovic, S. 2004. Birth and death: infant burials from Vlasac and Lepenski Vir.-Antiquity, 78: 301, 526-546.

Bronk Ramsey, C. 2009. Bayesian analysis of radiocarbon dates.-Radiocarbon, 51: 1, 337-360.

Chisholm, B. S., Nelson, D. E. & Schwarz, H. P. 1983. Dietary information from [[delta].sup.13]C and [[delta.].sup.15]N measurements on bone collagen.-14C and Archaeology: Symposium Held at Groningen, August 1981. Eds W. G. Mook & H. T. Waterbolk. (PACT, 8.) Council of Europe, Parliamentary Assembly, Strasbourg, 391-397.

Eisen, M. J. 1996. Esivanemate ohverdamised. Mats, Tallinn.

Finlay, N. 2000. Outside of life: traditions of infant burial in Ireland from cillin to cist.-World Archaeology, 31: 3, 407-422.

Jaanits, L., Laul, S., Lougas, V. & Tonisson, E. 1982. Eesti esiajalugu. ENSV Teaduste Akadeemia Ajaloo Instituut/Eesti Raamat, Tallinn.

Jonuks, T. 2002. Inimohver Eesti eelkristlikus usundis.-Maetagused, 19, 125-134.

Jonuks, T. 2009. Eesti muinasusund. (Dissertationes Archaeologiae Universitatis Tartuensis, 2.) Tartu Ulikooli Kirjastus.

Jurgenson, A. 1998. Tapetud laps ja lapsetapmine parimuses sotsiokultuurilisel taustal.-Maetagused, 8, 28-57.

Kalman, J. 2000a. Skeletal analysis of the graves of Kasekula, Poanse I and Poanse II.-Toid ajaloo alalt, II. Koost H. Pauts, toim T. Tamla. Eesti Ajaloomuuseum, Tallinn, 17-40.

Kalman, J. 2000b. Stone-grave II at Tougu-skeletal report.-Lang, V. 2000b, Appendix 3, 387-407.

Kalman, J. 2000c. Tandemagi stone grave-osteological report.-Lang, V. 2000b, Appendix 5, 423-436.

Kalman, J. undated. Joelahtme skeletal report. (Manuscript in the archive of the Institute of History at Tallinn University.)

Kraut, A. 1985. Die Steinkistengraber von Joelahtme.-TATU, 34: 4, 348-350.

Kriiska, A. & Saluaar, U. 1997. Aruanne arheoloogilistest valjakaevamistest Kasekula kiviaja asulakohal ja metalliaja kivikirstkalme aarevarel (Hanila khk.) 2.-28. juuni ja 10.-11. juuli 1997. (Manuscript in the archive of the Institute of History at Tallinn University.)

Kriiska, A., Lougas, L. & Saluaar, U. 1997. Archaeological excavations of the Stone Age settlement site and ruin of the stone cist grave of the Early Metal Age in Kasekula.-AVE, 1997, 30-43.

Lang, V. 1996. Muistne Ravala: muistised, kronoloogia ja maaviljelusliku asustuse kujunemine Loode-Eestis, eriti Pirita joe alamjooksu piirkonnas, I-II. (MT, 4. Toid arheoloogia alalt, 4.) Eesti Teaduste Akadeemia Ajaloo Instituut, Tallinn.

Lang, V. 2000a. Rescue excavations on fossil fields at Kasekula, west Estonia.-AVE, 1999, 70-78.

Lang, V. 2000b. Keskusest aaremaaks: viljelusmajandusliku asustuse kujunemine ja areng VihasooPalmse piirkonnas Virumaal. (MT, 7.) Tallinn.

Lang, V. 2007. The Bronze and Early Iron Ages in Estonia. (Estonian Archaeology, 3.) Tartu University Press.

Lang, V. 2011. Traceless death: missing burials in Bronze and Iron Age Estonia.-EJA, 15: 2, 109-129.

Lang, V., Laneman, M., Ilves, K. & Kalman, J. 2001. Fossil fields and stone-cist graves of Rebala revisited.-AVE, 2000, 34-47.

Laul, S. 2001. Rauaaja kultuuri kujunemine Eesti kaguosas (500 eKr-500 pKr). (MT, 9. Opetatud Eesti Seltsi Kirjad, 7.) Tallinn.

Lidstrom Holmberg, C. 1998. Prehistoric grinding tools as metaphorical traces of the past.-Current Swedish Archaeology, 6, 123-141.

Lougas, V. 1970. Sorve laevkalmed. - Studia archaeologica in memoriam Harri Moora. Toim M. Schmiedehelm et al. Eesti NSV Teaduste Akadeemia Ajaloo Instituut / Valgus, Tallinn, 111-118.

Lougas, V. 1971. Uber die Typologie und Chronologie der altesten Hirtenstabnadeln des Ostbaltikums. Suomen Museo, 78, 20-37.

Lougas, V. 1975. Uber die Entstehnung ortsgebundene Bodenbaukultur in Westestland.-TATU, 24: 1, 85-88.

Lougas, V. 1976. Ausgrabungen der Steingraber und Flurrelikte in Iru.-TATU, 25: 1, 48-52.

Lougas, V. 1984a. Uber die Forschung der Bodendenkmaler der Zeitwende auf den Westinseln Estlands.-TATU, 33: 4, 349-352.

Lougas, V. 1984b. Kopu Pihla kivikalmete uurimisest Hiiumaal. (Manuscript in the archive of the Institute of History at Tallinn University.)

Lougas, V. 1996. Kaali kraatrivaljal Phaethonit otsimas. Eesti Entsuklopeediakirjastus, Tallinn. Maldre, L. 2000. Tougu II kalme arheozooloogiline materjal.-Lang, V. 2000b, Appendix 4, 409-422.

Mandel, M. 1973. Aruanne Kasekula kivikirstkalme kaevamisest 1973. aastal. (Manuscript in the archive of the Estonian History Museum.)

Mandel, M. 1975. Ausgrabungen der Steingraber von Kasekula.-TATU, 24: 1, 74-76.

Mandel, M. 2000. Poanse tarandkalmed.-Toid ajaloo alalt, II. Koost H. Pauts, toim T. Tamla. Eesti Ajaloomuuseum, Tallinn, 89-111.

Mandel, M. 2003. Laanemaa 5.-13. sajandi kalmed. (Toid ajaloo alalt, 5.) Eesti Ajaloomuuseum, Tallinn.

Mandel, M. 2011. Arheoloogiga Laanemaa radadel. Eesti Ajaloomuuseum, Tallinn.

Murail, P., Maureille, B., Peresinotto, D. & Geus, F. 2004. An infant cemetery of the Classic Kerma period (1750-1500 BC, Island of Sai, Sudan).-Antiquity, 78: 300, 267-277.

National Registry of Cultural Monuments. [WWW] http://register.muinas.ee (28.03.2012).

Paulson, I. 1997. Vana eesti rahvausk: usundiloolisi esseid. Ilmamaa, Tartu.

Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., Buck, C. E., Burr, G. S., Edwards, R. L., Friedrich, M., Grootes, P. M., Guilderson, T. P., Hajdas, I., Heaton, T. J., Hogg, A. G., Hughen, K A., Kaiser, K F., Kromer, B., McCormac, F. G., Manning, S. W., Reimer, R. W., Richards, D. A., Southon, J. R., Talamo, S., Turney, C. S. M., van der Plicht, J. & Weyhenmeyer, C. E. 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0-50,000 years Cal BP.-Radiocarbon, 51: 4, 1111-1150.

Schmiedehelm, M. 1952. Rauaaja algus ja viljakasvatuse kujunemine pohiliseks elatusviisiks Kirde-Eesti alal.-TATU, 1: 3, 13-28.

Spreckelsen, A. 1927. Das Graberfeld Laakt (Lagedi), Kirchspiel St. Jurgens, Harrien, Estland. (Opetatud Eesti Seltsi Toimetised, XXIV.) C. Mattiesen, Dorpat.

Tauber, H. 1983. [sup.14]C dating of human beings in relation to dietary habits.-14C and Archaeology: Symposium Held at Groningen, August 1981. Eds W. G. Mook & H. T. Waterbolk. (PACT, 8.) Council of Europe, Parliamentary Assembly, Strasbourg, 365-375.

Thedeen, S. 2003. Life course practices in Bronze Age landscapes of East Central Sweden: beyond divine chiefs and neodiffusionism.-Current Swedish Archaeology, 11, 97-118.

Tvauri, A. 2012. The Migration Period, Pre-Viking Age, and Viking Age in Estonia. (Estonian Archaeology, 4.) Tartu University Press.

Valk, H. 2001. Rural Cemeteries of Southern Estonia 1225-1800 AD. 2nd edition. (CCC Papers, 3.) Gotland University College, Centre for Baltic Studies, Visby; University of Tartu, Archaeology Centre, Tartu.

Vassar, A. 1938. Drei Steinkistengraber aus Nordestland.-Opetatud Eesti Seltsi Aastaraamat, 1937, 1, 304-364.

Vassar, A. 1941. Kaevamisaruanne Jamaja khk. Torgu vl. Karuste kulas 1940. a. (Manuscript in the archive of the Institute of History at Tallinn University.)

Vassar, A. 1943. Nurmsi kivikalme Eestis ja tarandkalmete areng. Doktoritoo. Tartu Ulikool, Tartu. (Manuscript in the University of Tartu Library.)

Vedru, G. 1998. Kahala jarve umbruse asustuspiirkond muinasajal. Magistritoo arheoloogia erialal. Tartu Ulikool, Tartu. (Manuscript in the University of Tartu Library.)

(1) All the numbers of potsherds in this article are according to the excavation reports. Counting of the sherds today would result in a different number due to their continuous fragmentation in museum stores.

(2) The details of the find context for the third item (AM 476: 55) are unknown, except for the fact that it came from the north-western part of the grave (Mandel 1973). Also, it was not present in the museum at the time the finds were studied for this article. Excavation report claims this to be fragment of a grinding stone, which suggests that it may have been similar to the item in Fig. 4: 7.

(3) In this context a reference should be made to the isotopic values of bone samples (see Table 3), which apparently suggest a preference of the Kasekula population for terrestrial food resources (e.g. Chisholm et al. 1983; Tauber 1983), although in the Bronze Age a narrow cove lay not much farther than a kilometre to the south-east of the contemporaneous settlement. Also, a relatively high rate of dental caries observed on the teeth of the buried individuals indicates a general reliance on agriculture (Kalman 2000a, 21). An in-depth discussion of the isotope values and diet of the sample population is, however, excluded from this article.

(4) Double sampling was not entirely intentional, however. When collecting the samples, it was impossible to locate the bone assemblage No. 63 within the grave area, because the respective information was missing in the excavation records. So I decided to test if radiocarbon dating could determine whether these bones of a ca 50-year-old male belonged to a man in the central cist, to a man on the limestone paving north of the grave circle or even to a third person unrecorded by anthropologists. Only later Mati Mandel, referring to his excavation notes from 1973, kindly confirmed that the bones in question came from the central cist (Mandel, pers. comm. 27.03.2012). Radiocarbon dates agree with this statement.

(5) Given that the square for the both samples is the same (4/p) and that the cist has been extensively destroyed, for the sake of simplicity I describe the context of sample No. 4 as 'destroyed cist' in the tables and figures, even if the skeleton was actually located next to and not inside the cist.

(6) This bone deposit appears to be the only one that could be suspected to be a secondary burial, since other skeletons were either more complete or, if otherwise, severe damage or natural fragility of (infant) bones could be held responsible. However, the observed incompleteness of the skeleton could have resulted from a shallow grave, an explanation not less likely than the inference of a secondary burial.

(7) In view of the surprising results of Allmae's dating, it would have been worthwhile to take a few samples from the infant skeletons found in the central cist. Considering the high cost of AMS dating, this was not done, as AMS dates for the stone-cist graves at Rebala, Harjumaa (unpublished data in the possession of author) had suggested that contrary to the subadult burials outside the cists, subadult burials inside the cists are of the same date with the adult burials in the cists.

(8) In addition to the child, the respective bone assemblage (No. 40 in the excavation records) included the remains of an infant, which were radiocarbon-dated by Allmae (Table 2).

(9) For the period of 700 years (e.g. 900-200 BC), a prehistoric settlement unit with on average seven members expectedly produces roughly 200 dead (for the argument, see Lang 1996, 354 f.; 2007, 223). At Parnamagi this would make an average of roughly 20 burials per grave. This is not impossible theoretically, but it is still unlikely because this number of Early Metal Age burials would be unusually high for a stone-cist grave. On the other hand, the figure applies only if all the members of the community were interred in a stone grave-which need not have been the case and current opinions in the academia prefer the view that it indeed was not (e.g. Lang 1996, 354 f.; 2007, 223 ff.; cf. Lang 2011, 114 f.).

(10) It is in fact questionable even at Kasekula whether the Viking Age infant burial reportedly found in the central cist had deliberately been placed in the cist. Three thousand years after their construction, the cists of stone-cist graves are usually found to be full of soil and stones, and this may at least partly have been the case also a thousand years ago. People of the time need not have been aware of the presence of a stone cist, i.e. the infant in question may have been interred in the area of the cist unintentionally. It is impossible, however, to verify this hypothesis, because the infant burials were distinguished only during the later study of bones and no stratigraphic peculiarities were noticed at excavation (Mati Mandel, pers. comm. 25.03.2012).

Table 1. Comparison of the results of two osteological studies
(Kalman 2000a and Allmae 2010) of the human bone assemblage
from grave I at Kasekula

Context                 Kalman 2000a        Allmae 2010

Central cist         1 male, 50+ years   1 male, 50+ years
                     3 infants,          4 infants,
                     0-1 years           0-0.5 years
Outside the central  1 child, 3-1 years  1 child, 3-1 years
cist, within the     16 infants,         21 infants,
stone circle         0-1 years           0-0.5 years
Extension'           1 male, 50+ years   1 male, 50+ years
north of the         1 adult
stone circle

Total                23                  28

Table 2. Radiocarbon dates of human bones from grave I at
Kasekula by Raili Allmae (after Allmae 2010, Table 1). Calibration
by computer programme OxCal v4.1.7 (Bronk Ramsey 2009;
Reimer et al. 2009)

Context/square    Sex/      Bone /      Lab. no.
                  age    register no.   Poz-...

Central cist      Male   Cranial vault   32 412
                  50+         29
Central cist     Infant      Femur       32 413
                              29
Eastern part of  Infant  Temporal bone   32 414
the grave 9/q                 40

                   Date BP
                    ...         Date cal
Context/square   [+ or -] 30    (95.4%)

Central cist        2780      1005-840 BC
Central cist        1195       AD 720-940
Eastern part of      920      AD 1030-1185
the grave 9/q

Table 3. Radiocarbon dates and stable isotope measurements of the human
bones of grave I at Kasekula. Calibration after OxCal v4.1.7
(Bronk Ramsey 2009; Reimer et al. 2009). The uncertainty in the stable
isotope measurements is [+ or -]0.1[per thousand], unless otherwise
stated

Context/ square        Sex/ age   Bone/ register No.  Lab. No.
                                                      Hela-...

1 Central cist        Male 50+    Femur (?) 29          2418
(southern
and middle part)

2 Central cist        Male 50+    Distal phalange 63    2423

3 Destroyed           Adult       Fibula 50             2421
cist 4/p

4 Destroyed           Adult       Tubular bone N/A      2420
cist (?) 4/p

5 North of the        Male 50+    Tubular bone 41       2419
stone circle 5/p

6 SW part of          Adult       Unidentified 4        2422
the grave 10-11/m

7 Eastern part        Child 3-4   Humerus 40            2424
of the grave 9/q
(central part)

8 SE part of          Infant 0-1  Tubular bone 31       2426
the grave 11/p
(northern part)

9 Near the SW         Infant 0-1  Tibia 14              2425
corner of the cist
10/o (northern part)

Context/ square       Date BP ...    Date cal    [delta][sup.13]C
                      [+ or -] 30     (95.4%)

1 Central cist            2775      1000-840 BC        -20.1
(southern
and middle part)

2 Central cist            2801      1040-850 BC        -20.1

3 Destroyed               2728      930-810 BC         -20.2
cist 4/p

4 Destroyed               2573      810-570 BC         -20.0
cist (?) 4/p

5 North of the            2378      715-695 BC         -21.1
stone circle 5/p                    540-390 BC

6 SW part of              2089       200-40 BC         -20.8
the grave 10-11/m

7 Eastern part            2118      340-325 BC         -20.6
of the grave 9/q                     205-50 BC
(central part)

8 SE part of              2263      400-210 BC         -20.1
the grave 11/p
(northern part)

9 Near the SW             1073      AD 895-1020        -19.4
corner of the cist
10/o (northern part)

Context/ square       [delta][sup.15]N

1 Central cist              10.2
(southern
and middle part)

2 Central cist              10.0

3 Destroyed                  9.7
cist 4/p

4 Destroyed                  9.9
cist (?) 4/p

5 North of the               9.9
stone circle 5/p

6 SW part of                 9.3
the grave 10-11/m

7 Eastern part              10.6
of the grave 9/q
(central part)

8 SE part of                 8.5
the grave 11/p             [+ or -]
(northern part)              0.2

9 Near the SW               11.3
corner of the cist
10/o (northern part)