Printer Friendly

Archaeomagnetic studies of the Urartian civilization, eastern Turkey.

Introduction

Archaeomagnetism offers a method, established by Thellier (1938), for dating fired structures or objects by comparing the magnetization acquired as they originally cooled, with the properties of the Earth's magnetic field at that time. We present here the first archaeomagnetic study of Bronze Age sites in Eastern Turkey. As the geomagnetic field properties for this time and area are not known, these observations are correlated with observations in Bulgaria (Kovacheva 1980), the Ukraine (Rusakov & Zagniy 1973) and Iraq (Hammo-Yassi 1983).

Three localities were sampled -- Cavustepe (CA), Dilkaya (DI) and a cemetery next to Van Castle (VA). All samples were collected by the disk method (Tarling 1983); nonmagnetic disks are glued to the objects and then oriented, using a sun compass, before removing them with a small piece of the object, generally c. 8 cu. cm, attached to it. Incremental demagnetization of the remanence was undertaken on all but 5 samples using alternating magnetic fields (3, 5, 7.5, 10, 15, 20, 30, 40, 50, 60, 80 & 100 mT peak) in order to remove any magnetizations acquired subsequent to the last firing. Incremental thermal demagnetization (20, 50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550 & 600[degrees]C) was undertaken on 5 Cavustepe samples to determine whether their last firing temperature could be estimated and to ensure that the original single component of remanence had been satisfactorily isolated. In all sites, the magnetic remanences of the samples were measured using spinner magnetometers and each demagnetization vector was analysed independently for stability and principal components. The linear vectors were defined as having a diagonal error of [greater than or equal to]5[degrees] (Kirschvink 1980), although all linear components with diagonal errors of [greater than or equal to]10[degrees] were examined.

Cavustepe

This fortress, some 25 km south of Van at 38.27[degrees]N 43.50[degrees]E, was built on a pristine site on an 800-m high limestone ridge by the Urartian king Sarduri II (Sardun or Seduri in Assyrian texts) in the middle of the 8th century BC. It comprises two main units (FIGURE 1) separated by a col and has been excavated by Afif Erzen (1987). The upper, eastern and smaller part is largely a temple (D) dedicated to the god Haldi. The lower, western part of the fortress, consists of a palace complex (A) and a temple (B) dedicated to the god Irmusini. The palace, some 81 x 33 m, includes solid-rock water cisterns and an advanced plumbing system, including water supplies directly to the kitchen and a royal latrine. One of the several rooms (C) between the palace and the temple, c. 5 x 20 m, contained 52 large Urartian pottery storage jars, each some 2 m high; the room had been backfilled, leaving only the c. 1-m diameter rims exposed. These vessels contained large quantities of grain that had been burnt, presumably when the entire fortress was destroyed and burned by the Medes about 590 BC. Five of these storage jar rims were sampled in the granary storage area on the west (locality C on FIGURE 1), including symmetrical sampling (6--7 samples) of the rims of 3 vessels (CA1, 4 & 5). One fired clay vessel, also c. 1 m in diameter, was also sampled (CA7) on the higher eastern side, c. 50 m east of the entrance to the Haldi temple. This jar, also Urartian and c. 1 m in diameter, is of unknown purpose. Areas of natural limestone, reddened by the heating assumed to have occurred during the destruction of the entire site, were also sampled (CA6) at the entrance to the Irmusini temple. All sampling was undertaken on 26 August 1989.

[CHART OMITTED]

Results

Storage vessels (CA1--5)

(i) Alternating magnetic field demagnetization behavior

All samples behaved similarly to incremental demagnetization with very little sign of low coercivity (viscous) components; one clear vector was present in most samples (TABLE 1 & FIGURE 2b) although most isolated vectors did not include the initial natural remanence determination. In some instances, computer analyses (based on Kirschvink 1980) indicated two or more vectors; visual inspection showed these were parts of the same component and the vector corresponding to the widest coercivity spectrum has been taken as characteristic (TABLE 1). In most samples, the most stable direction, determined irrespective of intensity changes (Tarling & Symons 1967), was identical to that determined by linearity analysis; mean values, whether based on stability or linearity, were essentially identical (TABLE 2).

[TABULAR DATA OMITTED]

(ii) Thermal demagnetization behaviour

Two samples from CA1, two from CA5 and one from CA4 were thermally demagnetized. All behaved similarly, showing high stability and mostly revealing a single vector, although the initial natural remanence was only included within the vector for vessel CA4, suggesting that a very low blocking component (<50[degrees]C) was removed before the characteristic vector was isolated which was then present to at least 600[degrees]C, with no evidence of any other vectors (FIGURE 2a). This suggests that the samples had all acquired their present remanence while cooling from a single heating to [greater than or equal to]600[degrees]C. Both alternating and thermal demagnetization methods were mutually consistent. The similarity in the directions of the thermally isolated and magnetically isolated vectors further confirmed that single vector nature of the remanence after the removal of a minor viscous component carried by very low coercivity and blocking temperature grains. On this basis, the mean directions (TABLE 2) have been calculated using both thermally and magnetically isolated components.

[CHART OMITTED]

[TABULAR DATA OMITTED]

Conclusions concerning the storage vessels

When each pot was originally fired, it would have acquired a thermal remanence with a declination and inclination the same as the geomagnetic field within the kiln. When placed in position in the storage room, it is possible that the inclinations remained consistent (if all were originally fired upright or inverted) but their declinations would be scattered. It is clear (FIGURE 3a) that while the range in declination is quite large, this largely reflects the very steep inclinations, and the mean vectors for each pot lie entirely within the northeast quadrant. Thus all the observations suggest that their characteristic remanence is entirely related to the destructive burning at the site, i.e. at the time that their contents were carbonized.

[CHART OMITTED]

The mean direction for all pots is probably that of the geomagnetic field at that time as the directions from samples taken from around each rim are mutually consistent and there is no curvature of the magnetic vector during thermal demagnetisation; a feature which would be indicative of magnetic refraction effects (Tarling et al. 1986). The scatter is slightly larger than would be expected for such ideal materials, probably reflecting difficulty in obtaining precise orientation on curved, broken surfaces. It is concluded that the mean remanence direction, (decl = 34.7[degrees], incl = 73.5[degrees], [[alpha].sub.95] = 8.6[degrees]) is that of the geomagnetic field around 590 BC.

Separate storage vessel (CA7)

Only three samples were taken from this vessel; all showed a single component of high stability to alternating magnetic fields with identical inclinations, 45--46[degrees], for two samples although their declinations differed by 100[degrees] and the third sample vector was completely different. The mean direction of the two more consistent vectors (decl = 55[degrees], incl = 58[degrees]) is consistent with the directions isolated in the vessels within the storage room, but the scatter casts doubt on such comparisons.

Burnt limestone (CA6)

As the limestone was only weakly magnetized, the observed signal was unmeasurable after demagnetizing in peak fields of 7.5 to 20 mT. The vector isolated in one sample (decl = 83[degrees], incl = 66[degrees]) was broadly similar to that of the storage vessels but the weak intensity and directional scatter prevents further interpretation.

Dilkaya

This site, at 38.26[degrees]N, 43.13[degrees]E, 30 km southwest of Van (FIGURE 4), is thought to have been occupied around 2400 to 2000 BC (Cilingiroglu 1987 & pers. comm. 1991). Burnt mud bricks were sampled from a fireplace area on the north wall (DI 1--6) and from the burnt parts of the west wall (DI 7--9) of a room within a house complex, M5. Alternating field demagnetization revealed similar single components of remanence whether analysed for stability or linearity (TABLE 1) suggesting that only one heating event was recorded by each sample. Sample 2 disintegrated before measurement, but six of the other vectors were of similar orientation (FIGURE 3b) and the deviation of samples 6 (and possibly 5) was not surprising as their in situ nature was not completely clear in the field. The mean directions of remanence (decl = 355.6[degrees], incl = 45.0[degrees], [[alpha].sub.95] = 12.3[degrees] excluding 5 & 6; or 2.0[degrees], 44.8[degrees], [[alpha].sub.95] = 13.7[degrees] excluding sample 6) corresponds to geomagnetic field direction at the time that both the northern and western walls were burnt at the time of abandonment of this room.

[CHART OMITTED]

Van Castle cemetery

A single pot, 60 cm in diameter and of unknown height, was sampled in the cemetery area (38.50[degrees]N 43.26[degrees]E) to the north of Van Castle (Van Kalesi). The Castle itself dates from the time of Sardun (Seduri) I, who ruled between about 832 and 825 BC. The age and context of this pot is uncertain. As the final destruction of Van Castle is generally attributed to the Medes in about 590 BC, it seems reasonable to assume that the age of the pot is of the late 6th century BC. The direction of remanence (FIGURE 2d) in all four samples were single vectors and all in agreement (TABLES 1 & 2), but the mean direction does not bear any likely relationship with the probable direction of the geomagnetic field at this site. The vessel showed no signs of extensive secondary heating so the remanence was probably acquired while being fired. The very shallow inclination, relative to the rim, suggests that the pot was last fired while at an angle to the geomagnetic field, i.e. not standing on its rim or base. The direction of magnetization of this pot (274.5[degrees], 2.6[degrees]) cannot therefore usefully be used to determine the Earth's magnetic field direction; its firing orientation could be determined once an archaeomagnetic curve has been established.

Discussion

The fire destruction of Cavustepe resulted in very intense heating of the area. When cooling, the area acquired a remanence with a direction of the geomagnetic field (decl = 34.7[degrees], 73.5[degrees], [[alpha].sub.95] = 8.6[degrees]). This is much steeper than the direction associated with the axial geocentric dipole at this locality (0[degrees], 57.4[degrees]) and provides a reference point that enables testing for the synchroneity of other fired destructions in the region. It also constrains the geomagnetic secular variation for the time of the fortress destruction, currently considered as about 590 BC. The Dilkaya site similarly records the direction of the geomagnetic field at a specific, but as yet undefined time some 4000 to 4600 years ago. Until research provides a more precise date for this event, this observation cannot be used to define secular variation, although relative archaeomagnetic dating can be attempted for other destructions within southeast Europe.

Direct comparison with archaeomagnetic observations from Iraq, Bulgaria and the Ukraine (TABLE 3) is difficult because of the great distances involved and the uncertainty in the firing ages. As the inclined geocentric dipole correction is thought satisfactory (errors <1[degrees]) for areas within 600 km of the central point (Tarling 1983; 1989), this has been used (FIGURE 5). The mean Dilkaya inclination is in excellent agreement with those of the 20th and 21st centuries BC in Bulgaria, suggesting that the firing of Dilkaya was close to the 20th century BC, but the absence of other data until the 30th century BC restricts this conclusion. The high inclination for Cavustepe is supported by Iraqi observations for Babylon and Assur; the very different declination for Cavustepe is not considered significant in view of the steepness of the inclination. The archaeomagnetic data thus broadly supports the proposed date, around 590 BC, for the destruction of this site. Both archaeomagnetic observations are consistent with expected ages and initiate an archaeomagnetic database for Eastern Turkey.

[TABULAR DATA OMITTED]

[CHART OMITTED]

Acknowledgements. We are particularly grateful to Ersin Kavakli, the Director of the Archaeological Museum at Van, and particularly Muhittin Toprak, for their invaluable assistance in locating and permission to sample these sites. We are also grateful to Dr Aynur Ozfirat of the Liberal Arts Faculty of Istanbul University for her assistance and to A.L. Abdeldayem for helpful discussions. The result was funded by NATO grant No. 0341/88.

References

CILINGIROGLU, A. 1987. IX. Kazi sonuclari toplantisi I, Van--Dilkaya Hoyugu kazisi T.C., Kultur ve Turizm Bakanligi Eski Eserler ve Muzeler Genel Mudurlugu. Ankara.

HAMMO-YASSI, N.B. 1983. Archaeomagnetic work in Britain and Iraq. Ph.D thesis, University of Newcastle upon Tyne.

KIRSCHVINK. J.L. 1980. The least--squares line and plane and the analysis of palaeomagnetic data, Geophysical Journal of the Royal Astronomical Society 62: 699--718.

KOVACHEVA, M. 1980. Summarised results of the archaeomagnetic investigation of the geomagnetic field variation for the last 8000 yr in south-eastern Europe, Geophysical Journal of the Royal Astronomical Society 61: 57--64.

RUSAKOV, O.M. & G.F. ZAGNIY. 1973. Archaeomagnetic secular variation study in the Ukraine and Moldavia, Archaeometry 15: 153--7.

TARLING. D.H. 1983. Palaeomagnetism. London: Chapman & Hall.

1989. Geomagnetic secular variations in Britain during the last 2000 years, in F.J. Lowes et al. (ed.), Geomagnetism and Palaeomagnetism: 55--62. Dordrecht: Kluwer Academic.

TARLING D.H. & D.T.A. SYMONS. 1967. A stability index of remanence in palaeomagnetism, Geophysical Journal of the Royal Astronomical Society 12: 443--8.

TARLING, D.H., N.B. HAMMO & W.S. DOWNEY. 1986. The scatter of magnetic directions in archaeomagnetic studies, Geophysics 51: 634--9.

THELLIER, E. 1938. Sur l'aimantation des terres cuites et ses applications geophysiques. Annales de l'Institut de Physique du Globe 16: 157--302.
COPYRIGHT 1993 Antiquity Publications, Ltd.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Saribudak, M.; Tarling, D.H.
Publication:Antiquity
Date:Sep 1, 1993
Words:2346
Previous Article:Establishing weight systems in Bronze Age Scandinavia.
Next Article:Looting the antiquities of Mali: the story continues at Gao.
Topics:

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters