The Provenance of anthropomorphic Figurines found at Khirbat al-Mudayna in Moab as compared to similar Figurines found at Khirbet Qitmit in Israel, by Neutron Activation Analysis Jan Gunneweg and Marta Balla INTRODUCTION The study of pottery is usually one of the few windows we have into the lives of by-gone civilizations. Not only might one get a view into the evolution of an early technology based upon pottery production but also might one learn in some detail about the development of what may well be the earliest commercial trade or sound human interrelations between individuals or groups of people. This applies to all the sites in Israel and Jordan that will be treated in this paper. GOAL The present study is focused upon the establishment of the provenance of pottery found at the Horvat el-Mudayna road shrine in light of what has been found in the Horvat Qitmit and ‘En-Hazeva shrines together with pottery from Busayra. We made it our task to determine whether painted ‘Edomite’ ware that is easily discernible in Jordanian as well as in Israeli sites could have had its origin in Edom proper and whether Judah pottery would show up in Edom. If this were true, one might point to a form of commerce or just interrelations between the various sites and regions. ANALYTICAL TECHNIQUE The technique to trace pottery to the potter's workshop where it was made is based on the establishment of a chemical fingerprint of a ceramic by means of neutron activation analysis (hence NAA) that establishes the abundances of chemical elements in any earthy matter. NAA is based on the premise, after 35 years proven to be accurate, that each clay source on Earth has a different geo-chemical history, hence a different chemical composition and can, therefore, be distinguished from all others (Perlman & Asaro 1969; Gunneweg et al. 1994; Molnar, 1993:45-64). The fingerprint is based by large on trace elements that are site specific. The method has practically been unaltered since 1969 with only minor corrections for some of the chemical elements measurements in the Standard Pottery When potters add tempering materials to their clay, a new chemical composition is formed that is unique to that potter or to a site in general. NAA has the capability to tell also the ceramics made by these potters and of these clays apart. It determines quantitatively about 35 chemical elements-- for the major part trace elements-which are sufficient to provide the so-called chemical "fingerprint" of a ceramic vessel. By subjecting a set of samples to a given neutron flux, the stable isotopes of the various elements present in a ceramic are converted into radioactive ones, which emit gamma rays. These mono-energetic gamma rays will later identify the element, while the abundance of the element is determined by counting the gamma-photons that are emitted in a predetermined length of time. The modus operandi is that samples of pottery vessels or shards are taken to establish a local chemical profile of the site. The pottery powder obtained by a synthetic sapphire drill is placed into a plastic cup and registered. The powder is then weighed and sealed in a polyethylene vial. These vials are irradiated in the nuclear reactor of the Technical University of Budapest with a neutron flux of 2.4x10 12. The Budapest facility uses a Canberra HPGe Well-type Detector that is connected to a Canberra S100 Multi-channel Analyzer that stores the Gamma-spectrometric measurements. For the evaluation of spectra, SAMPO 90 software was used. Standardization was made by the so-called single comparator method, using gold as the comparator element. By irradiating the NBS 1633a Coal Fly Ash Standard Reference together with the samples, the accuracy of the measurements is controlled. We are convinced that in the light of the results on the pottery assemblages published by us over the past 30 years, laboratory techniques can really be of assistance, also or perhaps specifically in the case of Edom-Moab -Judah sites because of their geographical dispersion. STATISTICS In our case, having to deal with a limited amount of pottery from various sites, geographically quite isolated from one another and also typologically different from pottery styles in other Edomite sites, the “goodness of fit” value Chi-Square to a given composition coupled with the Euclidean Distance measurement and an additional cluster program will suffice. Such a procedure as described by Mommsen (Mommsen et al. 1988) takes also care of experimental errors and a possible dilution phenomenon introduced by potter’s practices. On other occasions, such as the NAA study on ‘En Hazeva pottery, Principal Component Analysis (PCA) was used to separate chemical fingerprints that are near one another. The process has been described by L. Balazs as ‘Data Mining’ in Gunneweg and Balla PREVIOUS RESULTS. The first detailed study on the provenance of ‘Edomite’ pottery started in 1988 at the Hebrew University of Jerusalem and the Kern Physik Institut at Bonn on cult vessels found in the Horvat Qitmit shrine. Later, this research was enlarged into a collaboration project with the Nuclear Reactor Facility at the Technical University of Budapest (Balla and Gunneweg, in press).> The collaborative NAA study established the provenance of a representative assemblage of ‘Edomite’ and local pottery at Horvat Qitmit. The results showed that all figurines were locally made somewhere in the triangle between Beer Sheba-Arad and Horvat Qitmit in the North-eastern Negev of Israel (Gunneweg et al. 1991: Tables 2 and 4; Gunneweg and Mommsen 1995). Hence, it was decided that the anthropomorphic and zoomorphic figurines couldn’t be called "Edomite" in the strict sense as archaeologists had previously suggested. Previous samples from Edom, analyzed and published in the Bonn-Jerusalem collaboration program, consisted of three Edomite cooking pots obtained from Tel el Ghrareh in Jordan, which were traced to the Tawilan district. A cream ware bowl found at Busayra remained of unknown provenance. A painted jug found in Busayra, however, came from the ‘Umm el Biyara/ Petra region (Gunneweg and Mommsen 1991). Simultaneously, it was found in the then existing literature about southern Jordan that figurines of similar shape as found in the Qitmit shrine are almost entirely lacking in Edom where they originally were thought to come from. Ten years later in 1999, the research continued by looking at the provenance of a series of about forty-two cultic vessels and three anthropomorphic vessels that had been found in yet another shrine, at 'En Hazeva. The latter is located in the Arava, a desert between Israel and Jordan. About 45 samples were analyzed by NAA at Budapest (Balla and Gunneweg, in press). > When the former research was still going on, we received from P. Bienkowski some Edomite pottery from Busayra, biblical Bosrah (Gen 36:33). The site is one of the major cities located in the northern Edom Highlands (at 1500 meters height), in a straight line to the East of ‘En Hazeva in Wadi Arava.> From the results obtained at Busayra we were able to form a chemical group of pottery that matched pottery that must be deemed to be local to the site where it was found. First, the samples represented the most common types of pottery found at Busayra so that archaeologically speaking; they are local to the site. The painted shards were added to the sampling because we were interested to see whether earlier work done on painted “Edomite” pottery would provide similar results. The local Busayra group consisted of seven samples in all. A group of 12 samples from ‘Umm el-Biyara was also analyzed. The results will not be mentioned except for the fact that they are entirely different from those analyzed from Busayra. This chemical composition has not been found matching any of the other Jordanian sites, such as ‘Umm el-Biyara, Tawilan and Petra, so that we are inclined to call this group the local chemical fingerprint of Busayra. Three samples from Busayra were taken from painted vessels and they match the group that we have called ‘local Busayra’. It proves that the painted version of Busayra pottery is locally made whereas other painted ware was proved to have originated in ‘Umm el-Biyara. Five Busayra samples match chemically pottery groups established to have originated in ‘Umm el Biyara and Petra. Also the new samples freshly obtained from ‘Umm el-Biyara (Bienkowski’s collection) and analyzed at Budapest corroborate this statement. The overall conclusion is that we were able to compose a group of pottery that was found at Busayra and that probably was locally made there. On the other side, we have three vessels that surely came from the Petra area, which means that there were relations between Busayra and the ‘Umm el-Biyara (Petra) area. Is there pottery that originated in Busayra and that ended up in a Judahite or an Arava site, to the West of Jordan? Yes, when four vessels found in ‘En Hazeva, HAZ 6 (a zoomorphic vessel), 19, 22 and 81 (a bowl) individually are used as an input for obtaining the Chi-Dissimilarity and the Euclidean Distance, they score a number between 0-1 as postulated compared to BUS 8, 12 and 17 which belong to the local Busayra chemical group. This means that these vessels found in the Arava came from Busayra. BUS 5, a white bowl, is a sample that stands alone among the Busayra samples. However, it was found to be chemically similar to HAZ 102, a cultic stand. HAZ 102 was established to be local to ‘En Hazeva, so it may be an indication that also BUS 5 originated in the Arava where ‘En Hazeva is located. It seems that a pottery vendor or trader supplied Busayra with ‘En Hazeva pottery, or that either a traveler from Busayra or ‘En Hazeva chose this vessel and brought it to Busayra. The outcome of Busayra NAA results proves the thesis that each city in Edom, indeed, made its own ceramics. We were able to get a chemical fingerprint for Busayra as well as a confirmation of the ‘Umm el-Biyara pottery as described in some earlier publications. Shortly thereafter, it came to our attention that Michele Daviau had found in her excavation at Khirbat al-Mudayna in Moab, Jordan, a road shrine just beside a main north-south road between Madaba and Umm ar-Rasas with a group of broken figurines similar to those at Qitmit. Also this pottery belongs to the Late Iron Age II. Specific figurines were called "Moabite" by Daviau, whereas another style of ceramic figurines unearthed reminded her of those found in Khirbet Qitmit and 'En Hazeva (Daviau 1997, p.225). Eleven samples of ceramics found at Mudayna, among which the purported Qitmit-looking figurines were submitted to NAA at Budapest. Again, we wanted to know whether there is any connection between the figurines found at Horvat Qitmit, 'En Hazeva and now Khirbet al-Mudayna and whether one is able to trace their provenance by making use of the databanks of Bonn, Jerusalem and Budapest where Jordanian pottery is present. List of Moabite pottery samples. Moab 1 WT 13/A2 white ware pottery Moab 2 WT 13/A2 brown ware pottery Moab 3 WT 13/A2 pink ware pottery Moab 4 WT 13/A2 red ware pottery Moab 5 MT 061 MT/L 34-5-9 Moab 6 WT 13/A2 WT 165 Moab 7 WT 13/A2 WT 166 L.37 Moab 8 WT 13/A2-15-4 L.11 WT 168 Moab 9 WT 13/A2-3-7 L.2 WT 169 Moab 10 WT 13-RS WT/015 Pail-15 Moab 11 WT 13/A2 L.14 WT 171 When the NAA results came in, we looked for the characteristics whereby the various ceramics differ from one another and from NAA data of pottery already analyzed at other occasions. We noticed that the characteristics of the chemical composition of Moabite pottery could be summed up as follows: Barium is higher in samples 6 and 8 (almost 4000 ppm) than in the remaining samples where Ba is a factor of 10 less. We have, however, no clue why this is so. Calcium is high too; its abundance is more than 20 percent, except for samples 3, 4, 5 and 11. There is also an unusual high abundance of La in samples 6 and 8, i.e. 67 and 76 ppm, the same samples with the high Ba. In fact, all the rare earths in samples 6 & 8 are higher than in all remaining samples with Yb in sample 6 the highest (8.4 ppm). Uranium is another element with e.g. extremely high (32-70 ppm) values for samples 1, 2, 7 9 and 10, whereas medium high in samples 5, 6, 8 and 11 (11-13 ppm). U values in samples 3 and 4 are as normally is encountered in pottery (2-4 ppm.). We have not found a satisfactory answer to explain the high uranium values. If we compare this with what we have on our databanks, it becomes clear that the Moabite pottery is entirely different from that in Edom and Judah. By using the Chi-Square Dissimilarity coupled with the Euclidean Distance measurements as incorporated in Mommsen et al. 1988, as well as by Cluster Analysis at Budapest, the combined result was that samples Moab 1,2,7 and 9 are similar to each other and are not found in any chemical fingerprint of pottery on our data banks. Basing ourselves on these few samples we may assume that they were local to Moab. The chemical compositions of samples Moab 5 and 11 belong certainly together and could well be another fingerprint in Moab for which we did not find a match. Moab 6 and 8 are similar to an oil lamp found in 'En Hazeva that has been found in the Wadi Arava. Finally, Moab 3 has a chemical similarity to Busayra 8 and 12 that are painted bowls and 'En Hazeva 19 and 23, which are bowls with a black slip on which red concentric painted circles. The latter two from 'En Hazeva are "Edomite" vessels classified in the Hazeva paper as chemical group II, the "Edom connection". This means that also Moab 3 comes from Edom. Two samples, Moab 4 and 10 are chemical outliers. CONCLUSION The first piece of information is that although the figurines are stylistically similar to those found at Horvat Qitmit and 'En Hazeva, there is absolutely no connection between these sets of anthropomorphic and zoomorphic ceramic figurines found in these shrines and those in Khirbet al-Mudayna in Moab. There is an indication that Moab made its own figurines that stylistically are similar to those in Israel but were locally manufactured in Moab. To be fair, we have no way of knowing whether they are Moabite instead of being Ammonite but this has to wait until also Ammonite pottery is checked by NAA. Daviau stated in her paper that stylistically the Ammonite and Moabite figurines are different what makes us tentatively conclude that the Moabite pottery analyzed by us is indeed Moabite until more extensive tests have been performed. A single sample, Moab 3, analyzes as Edom local ware that also has been found in 'En Hazeva and Busayra and was traced to Edom. One thing is becoming clearer every time NAA is performed on Jordanian pottery, there is much more going on than the usual generalization by archaeologists of an Edomite presence or occupation of Judah and the Negev district. It has more the appearance of a trade that went on between the various peoples that occupied Jordan and Israel during the Late Iron II Age. Even when there were hostilities as described in the Bible, commerce did not stop. What concerns the statue that was found at Tell Boueid in Syria (Suleiman 1995), which has a great similarity to those at Horvat Qitmit and Hazeva, we will have to wait until NAA data will be available. REFERENCES Balla, M. and Gunneweg, J. The Search for the Manufacture Center of Cult Vessels Found in the Iron Age Shrine at 'En Hazeva in the Arava of Israel, (by Instrumental Neutron Activation Analysis) English ‘Atiqot Series, in press Beit Arieh, I. 1995, Horvat Qitmit, an Edomite Shrine in the Biblical Negev, Monograph Series of the Institute of Archaeology, Tel Aviv University Bienkowski, P. 2002, Busayra, Excavation by Cristal-M. Bennett 1971-1980, British Academy Monographs in Archaeology No. 13, Oxford University Press Daviau, M. 1997, Moab's Northern Border, Biblical Archaeologist, no 60, p.225 Gunneweg, J. and Mommsen, H. 1990, Instrumental Neutron Activation Analysis and the origin of some cult objects and Edomite vessels from the Horvat Qitmit Shrine. Archaeometry 32:7-18 Qitmit Gunneweg, J. Beier, Th., Diehl, U. Lambrecht, D. and Mommsen, H. 1991, 'Edomite', Negebite' and 'Midianite' Pottery from the Negev Desert and Jordan: Instrumental Neutron Activation Analysis Results, Archaeometry 33:239-253 Gunneweg, J. and Mommsen, H. 1995, Instrumental Neutron Activation Analysis of vessels and Cult Objects, 280-287 in Beit Arieh, Ed: Horvat Qitmit, an Edomite Shrine in the Biblical Negev, Monograph Series of the Institute of Archaeology, Tel Aviv University Gunneweg, J., Perlman, I. and Asaro, F. 1988, The Origin, Classification and Chronology of Nabataean Painted Fine Ware, Jahrbuch des roemisch-germanischen Zentralmuseums Mainz, 35, 315-345 Gunneweg, J. Asaro, F., Michel, H. and Perlman, I 1994, Interregional Contacts between Tell en-Nasbeh and littoral Philistine Centers in Canaan during Early Iron Age I, Archaeometry 36:227-239 Gunneweg, J. and Balla, M. 2002, Instrumental Neutron Activation Analysis, Busayra and Judah. In Bienkowski, Ed., Busayra, Excavation by Cristal-M. Bennett 1971-1980, British Academy Monographs in Archaeology No. 13, Oxford University Press, 483-86. Gunneweg, J and Balla, M. (in press) Neutron Activation Analysis, Scroll Jars and Common Ware, In: Khirbet Qumran and Ain Feshkha, Eds Humbert and Gunneweg,Vol. II., Editions Universitaires Fribourg Suisse, Vandenhoeck & Ruprecht Goettingen, 3-58 Molnar, Zs, Keoemley, G. Bodizs, D. & Lengyel, Z. 1993, Application of Neutron Activation Analysis in the Institute of Nuclear techniques of the Technical University of Budapest, in Periodica Polytechnica Series Physics, Vol.I, No. 1, pp. 45-64 Mommsen, H., Kreuser, A. and Weber, J. 1988, A method for grouping Pottery by Chemical Composition, Archaeometry 30:47-57 Perlman, I. and Asaro, F. 1969 Pottery Analysis by Neutron Activation, Archaeometry 11, 21-52 Suleiman, A. 1995, Tell Boueid. Chronique arch?ologique. Syria 72: p. 164-170. ACKNOWLEDGEMENTS Our sincere thanks to all the archaeologists who have let us sample the pottery treasted in the present study; I. Beit Arieh, P. Bienkovsky , R. Cohen and Y. Israel, M. Daviau and S. Hart and the late Mrs. C-M Bennett. Thanks to Z. Molnar at the nuclear reactor in Budapest for her expert handling of the analyses and L. Balazs for statistics throughout. APPENDIX List of individual Moab NAA data MOAB ITE SAMPLES Concentrations in ppm 1 2 3 4 5 6 7 8 As 9,7+-0,3 18,1+-0,4 6,9+-0,2 6,7+-0,3 10,2+-0,2 5,9+-0,2 20,9+-0,2 5,5+-0,2 Ba 1100+-100 1450+-130 1400+-100 1400+-100 1640+-80 2900+-100 760+-60 4100+-100 Ca 24,33+-0,24 27,61+-0,27 13,55+-0,14 12,40+-0,12 14,15+-0,15 22,04+-0,22 22,03+-0,23 18,04+-0,38 Ce 82,0+-2,0 70,2+-2,1 70,5+-0,7 43,6+-1,3 Xx 80,4+-0,8 66,0+-1,0 88,1+-0,9 Co 27,0+-0,5 21,7+-0,7 41,5+-0,8 13,6+-0,6 24,4+-0,4 23,2+-0,5 11,7+-0,2 22,8+-0,5 Cr 390+-4 396+-8 197+-6 136+-4 336+-5 217+-3 410+-5 237+-7 Cs Xx 1,6+-0,2 2,61+-0,37 2,50+-0,36 xx 1,88+-0,19 1,14+-0,18 1,63+-0,20 Eu 1,59+-0,11 1,64+-0,11 1,65+-0,10 0,91+-0,09 1,76+-0,07 2,89+-0,09 1,59+-0,06 3,14+-0,09 Fe% 3,59+-0,07 3,33+-0,07 6,53+-0,07 4,64+-0,05 5,93+-0,06 3,03+-0,03 3,24+-0,03 3,01+-0,03 Hf 5,61+-0,178 4,62+-0,32 4,73+-0,28 4,29+-0,30 6,72+-0,20 5,30+-0,21 5,50+-0,11 7,37+-0,22 K% xx xx 1,36+-0,13 1,84+-0,15 2,08+-0,10 xx 0,73+-0,10 Xx La 40,7+-0,4 39,7+-0,4 32,9+-0,3 19,2+-0,2 32,6+-0,3 67,3+-0,7 37,2+-0,4 76,6+-0,8 Lu 0,60+-0,02 0,61+-0,02 0,43+-0,02 0,29+-0,01 0,50+-0,01 1,26+-0,01 0,71+-0,01 1,43+-0,03 Na% 0,67+-0,06 0,43+-0,01 0,37+-0,01 0,27+-0,01 1,03+-0,01 0,36+-0,01 0,56+-0,01 0,66+-0,06 Nd 30+-8 Xx 28+-8 26+-6 28+-6 57+-5 30+-5 74+-7 Rb xx xx 48+-10 60+-10 Xx 36+-5 28+-5 33+-5 Sb xx 1,76+-0,04 0,72+-0,06 0,79+-0,04 1,27+-0,04 0,91+-0,03 2,66+-0,03 0,73+-0,02 Sc 13,6+-0,2 12,4+-0,12 20,8+-0,2 16,0+-0,2 20,8+-0,2 14,1+-0,14 12,9+-0,2 15,3+-0,2 Sm 10,3+-0,2 7,66+-0,15 6,68+-0,07 4,15+-0,04 6,71+-0,07 11,71+-0,12 8,43+-0,08 13,12+-0,13 Ta 1,09+-0,15 1,38+-0,16 1,41+-0,16 1,12+-0,16 1,33+-0,12 0,95+-0,09 0,77+-0,09 0,95+-0,09 Tb 0,99+-0,19 Xx 1,2+-0,2 xx 1,03+-0,14 1,98+-0,12 1,09+-0,12 2,26+-0,14 Th 7,3+-0,3 6,5+-0,3 7,7+-0,4 7,2+-0,3 7,4+-0,2 7,60+-0,15 6,71+-0,20 8,24+-0,08 U 70+-1 35,0+-0,7 4,47+-0,13 5,72+-0,17 12,48+-0,12 13,34+-0,27 48,7+-0,5 11,48+-0,23 Yb 4,16+-0,24 3,98+-0,16 3,17+-0,13 2,03+-0,04 3,76+-0,08 8,43+-0,17 4,61+-0,05 9,62+-0,10 Zn 326+-10 450+-10 176+-9 150+-10 249+-9 210+-10 360+-10 210+-10 MOAB ITE SAMPLES Concentrations in ppm 9 10 11 As 15,4+-0,3 17,5+-0,4 9,2+-0,3 Ba 1300+-70 1220+-40 1060+-90 Ca 24,81+-0,25 26,76+-0,27 15,43+-0,31 Ce 63,1+-1,3 41,7+-0,8 61,4+-1,2 Co 14,6+-0,3 7,66+-0,23 19,4+-0,4 Cr 363+-5 520+-5 319+-3 Cs 0,86+-0,09 0,63+-0,07 1,72+-0,22 Eu 1,54+-0,08 0,99+-0,05 1,63+-0,08 Fe% 3,33+-0,03 2,21+-0,02 5,07+-0,05 Hf 5,26+-0,21 2,89+-0,09 5,994+-0,18 K% xx xx xx La 37,87+-0,38 24,67+-0,25 30,47+-0,37 Lu 0,55+-0,02 0,51+-0,02 0,49+-0,01 Na% 0,51+-0,01 0,18+-0,01 0,68+-0,02 Nd 31+-6 26+-3 34+-5 Rb xx xx xx Sb 1,33+-0,05 1,99+-0,04 0,99+-0,08 Sc 11,5+-0,1 9,1+-0,1 18,9+-0,2 Sm 6,97+-0,14 5,44+-0,11 6,64+-0,07 Ta 0,94+-0,09 0,35+-0,07 1,23+-0,11 Tb 0,90+-0,10 0,65+-0,08 1,08+-0,14 Th 6,63+-0,13 3,68+-0,11 6,25+-0,19 U 24,42+-0,24 31,84+-0,32 11,30+-0,23 Yb 3,95+-0,08 3,26+-0,07 3,43+-0,10 Zn 403+-8 306+-6 285+-9 Comments? Please write: Jan Gunneweg To return, click Archaeometry abstracts of our work or click "BACK" or click the Homepage . Jan Gunneweg Ph.D., The Hebrew University, revised November 2005