DDendrochronological Dating in Anatolia: The Second Mileniurm BC Peter |. Kuniholm, Maryanne W. Newton, Carol B. Griggs & Pamela J. Sullivan Dendrochronological Dating in Anatolia: The Second Millennium BC In contrast fo the sketchy summary of dendrochronologi- cal information for the third and earlier milennia presented at the previous Bochum meeting (for a listing of sites and disconnected wiggle-matched dates see Kunihoim 1996), we report here a complete, robust, and continuous tree~ ring chronology for the second millennium BC +4/-7 years (Kromer et al. 2001; Manning etal. 2001). The error margin may actually be slightly lower (see Manning et al. 2003), and see now Figures 1 and 2. Al BC dates published here supersede those reported in earlier years. The relative dates remain the same. Figure ‘Aegean Dendrochronology Project (ADP) Bronze-Iron Master Chronology as of AD 2003, shown in terms of the 20-year moving average of the percent variation in ring Widths around normal (defined as 100) from all constituent data by year (the ‘Index Values’ ~ grey line). The number of securely cross-dated samples, an average of 32 trees peer year, which comprise this chronology is shown by the black line. The calendar date scale shown is the near-ab- solute dating proposed in Manning et al. (2001). For the specific trees from this chronology employed in the '*C \wiggle-match dating, see Figure 2. Although sample num- bers are not especialy large in the mid-16" century BC, we note that for the “C wiggle-match we employed a long- lived tree, GOR-161 with 861 tree-rings, which grew from the 18-10" centuries BC, It is securely cross-dated on the earty end against dozens of juniper trees from Porsuk (Kunihoim etal. 1992 and on-going work since), and then against, progressively, dozens, scores, and finally over 4100 trees from Gordion and environs. In addition to the data summarised above, newly developed uniper and pine toe a8 moun im 71s 90 nan man RRB aed 59 9 ior 61969 Rom Bon Mw uM Om ius 98 na isis 7 WH as. a 2 2 2 3 2 ist 1010 90 1078 tor 7s am 3 25 2 6 26 6 2 28 ma ust as 91s 357 058 99 non 7 nv» 0 0» 85 oll ass 90 106 1064736 Mo NOM Bw BH 3s 3s mm 30 ne 1008 B46 62 ae BOM 36 97 7 6 Me 36 mom rast 9s M50 M692 » wh Oba ws rd eos 976 6 045722 en 7 7H ww 0 0 mm 30 zt 76S 18 2M HSS ST 7 se ou nes as ee vow 12% 37 67 SR net ast bss 119 ma 1 oS 8 89 9 Oe nar 13s 10s 108t 0 “7 @oanaa sa o ome 1135 os a a Pn 8 8 @ 6 & aH HT ost aD 100 957 “ 28 RB WB TT HD iam ams ow 7 om ™ 2B HMB AG 8 9m cos ms oe os 6 a mwa 73988 103s as DOR 7% 8% 96 97 Ow os 953 mun m2 oe oR 5k oR oR oR OF set 126 6 968 64 = 98 9 98 100 105 10H vost toss ad ea > 406 106 105 108 108-108 08 ost 87 106s 113s 4 sow 107 109 166 106 107 108 107 to 1050 719 106 10s wr 107 105 105 104 104 105 303 M087 isk 86 1080 02 102 102 101 100 100 100 000721988 toe 102s ior on 100 59 99 100 100 100 101 so 1288 1010 st 36 foo 10 110 190 00 97 97 97 98 92 ‘Number of samples in data set: 284 ‘Number of rings in dataset: 35484 Length of dataset: 1081 years “Table 1: Supplement to Figure 1 For readers who wish to study the second millensium BC Aegean year-by-year, we provide @ set of tree-groth indices for the en- tire period, plus some 50 extra years on the ends. The small error margins are as noted in paragraph one of he text Instructions fr reading the tabe: Information forthe 1051 years from 2090 BC to 980 BC is presented as a growth index for each year (the lft column, ten years to a line numbered 0-1, and a histogram (the right column) show:ng the sample abundance for each year. The indices include a (men fa) decimal point, Thus the information for 1957 BC (which reads 775) should be understood as follows: average ring-growth was 7.5 % of normal (‘normal isthe mean growth index forthe period 1847-1967 BC), and this value isthe average annual growth in- ox of 18 diferent trees i a ho, cry yea. Similar, the information for 1648 BC (which reads 2070) shouldbe read as 207.0% of rormal, an average derved from 40 diferent tres In an extraordinarily cool, wet year. These 10S1 years are pat of a continuous 2009 year eequence which runs trom 2657 BC to 649 BC.Peter, Kuniholm, Maryanne W. Newton, Carl 8. Griggs & Pam la J. Sulivan we lack a continuous chronology through the ADIBC transi tion to alow us to assign absolute yeas, tis ating chrono- logy was pinned down by 68 C dates by Berd Kromer at Heidelberg as par of Sturt Manning's Eastem Mediterranean Radiocarbon Calibration Project (EMRCP). preliminary attempt to place the tree-ring chronology se- ‘cure in time on the basis of some 22 determinations was reported in Nature seven years ago (Kuniholm et al. 1996). Since then Maryanne Newton caught the steep downturn in the radiocarbon graph in the 8th century BC, getting us away from the relatively flat radiocarbon plateau of the centuries preceding the mid-8" century BC. Moreover, the discovery of a radiocarbon offset in the 9¥78" cen- turies BC, an apparent period of global cooling rather like the Little Ioe Age in the 15" and 16° centuries AD, was reported on in the December 21st, 2001 issue of Science (Kromer et al. 2004; Manning et al. 2001; Reimer 2001) and demonstrated the need for an upward shift of some 22 years from the placement we had offered five years eatlier. Accordingly, all the construction dates from the Middle and Late Bronze Ages and Eary Iron Age that we have been quoting over the years have to be raised by 22 years (+4/-7). ‘As a further test of this new placement, Bernd Kromer tried to catch the 1325 BC blip in the radiocarbon curve with six small segments ofthe Gordion wood, and as may be seen in the dotted circle in Figure 2 (Manning et al 2003) he has done so successfully. We cannot offer a {quantification yet of what this might mean for a modified precision range. For the time being we are holding to +4/- 7 years, but the error may be somewhat lower." Figure High-precision radiocarbon data, including six new data (Hd-21711, 21712, 21721, 21722, 21761, 21774) centered around the 1325 BC ‘wiggle’ in the radiocarbon calibra- tion curve, from 1O-ring samples of the Aegean Den- drochronology Project Bronze-Iron tree-ring series (Figure 1). The data are shown at thelr best-fit placement against the current (AD2002) internationally recommended INT- CAL9B radiocarbon calibration data set (Stuiver et al 1998). Samples were taken from three of the constituent trees of the well-eplicated Gordion-area dendrochrono- logy forming one of the ADP's longest floating sequences for the prehistoric Mediterranean and Near East. All ra- diocarbon measurements were made at the Heidelberg radiocarbon laboratory (see Kromer et al. 2001; Manning et al. 2001 for details). All data shown with 1®'errors. (Fi- ‘gure courtesy 8. W. Manning, See Manning et al. 2003 for further discussion). Lacking a bridge through Roman times to ving trees of the present, this Is as close to precision as we could se- riously hope for (in fact, it is rather better). An error mar- 44 Relative Ring Number 000 1200 14001600 Nn ay Svea a aS ow rovers Tinos nance ‘Shel ana (yarn he ‘iy 8 000 / ‘py, « nino no) 4 2°°° § mo Sat sre Ti eeeee= YL - Calendar Date BC gin of +4/-7 years is not all that bad for three to four thou- sand years ago. The reader will have noted four or five big peaks or bumps in Figure 1 on the histogram (dark line) of sample tre- quency, followed by a steep drop on the right-hand side of the bump, indicating major building activity atthe sites from which these samples were collected. Let us look at them from right to left, or from late to early, to see what they might mean, Bump 1 (ca. 673 BC) - Ayanis “The Haldi-Temple of Urartian King Rusa II was builtin 673, +4/-7 cal BC, rather early in Rusa’s reign which might ox- plain the lack of significant accomplishments with which he usually embellshes his inscriptions (and which at the time of excavation puzzled the epigraphers). There is no evidence of earlier or later construction at the site, so Aya- nis seems to have been a one-period affair. See Ci iroglu & Salvini (2001) and contribution by Kuniholm & Newton (200; 2002) therein for extended discussion. Note that the publication of this book preceded the dis- covery of our need to raise the date (Kuniholm & Newton forthcoming). Bump 2 (ca. 740 BC) - Gor ‘The so-called MM Tumulus at Gordion, built around 740, +4/-7 cal BC, 22 years earlier than we reported in 1996 in Nature, is therefore not the tomb of the quasi-historical ‘Midas who flourished and died around 700 BC but rather the tomb of someone a generation or two earlier. This find- ing has helped force a re-examination of the entire Gor- nDendrochronological Dating in Anatolia: The Second Millennium BC dion City Mound stratigraphic sequence. The June 2003 Antiquity (De Vries etal. 2003) has the latest restatement of the implications of this date for Gordion and Iron Age Anatolia. What had been thought at Gordion forthe last 50 years to be the Kimmerian destruction level ofthe ear- ly 7™ century BC we now realize is an otherwise-undocu- ‘mented early Iron Age destruction of the late 9"century BC. The dendrachronological date of 883 forthe joists of “Terrace Building 2A (but no bark visible and therefore pos- sibly slightly later) is complemented by radiocarbon tests Cf the thatch of the roof and of a variety of seeds found in separate pots in the destruction debris (the latter all Clustering between 830-800), also by ceramic links with imported Greek pottery in the levels above the burned lay- It there ever was @ Kimmerian attack on Gordion, it was ‘on the upper buildings above the clay which for years were called the ‘Persian’ buildings and are now more property termed Middle Phrygian. Early ron Age buildings at Gor: don such as Megarons § and 6 are 10" century (speci fically 940 and 944 +4/-7 cal BC), and so forth Bump 3 (circa 1549 BC) - Porsuk At this Hite site just north of the entrance to the Cilcian Gates to the Taurus Mountains the timbers in the substrate of the postem gate are much earlier than what has been deemed to be 14°/13" century pottery found in the de> struction debris above, The junipers, cedars, and pines found under the floor of the poster gate on the west side ‘of the mound were all cut in 1849 +4\-7 cal BC. An inner part of the gate was built 1 years eariier(Kuniholm et al. 11992; Pelon 1982) ‘An oddity at Porsuk is an enormous spike in growth at ca 11850 BC. The spike occurs in 61 out of 6! junipers, cedars, and pines, ranging in age ftom 19 to 244 years, and re- flects a spring/summer growing season that was extraor- dinarily cool and moist. This Is the most remarkable such anomaly in the last 9000 years, and we think itis a re- action to the eruption of Thera/ Santorini some 820 kms to the west.The spikes taper off a year or two or three later, andthe trees resume their normal lives until they are cut down. See Hammer (2003) for the latest statement about the Greenland ice core evidence and possible link- ‘age with the Porsuk evidence. Bump 4 (ca. 1774 BC) - Acemhéyiik: Sarikaya Palace and Hatipler Tepe: In these two Middle Bronze Age buildings, both of them but in the same year, we have quantities of burned logs, with the bark preserved. The last rings from timbers in both buildings cluster at 1774 +4)-7 cal BC. That is the construction date, Asinglerepairlogin the Sankaya Palace has a rng preserved from 1766 44-7 cal BC with an un- known number of rings burned of. So the building had at least an eightyearIte-span before the fe. A full report on these two important monuments has yee appear. For prolimary reports see Oz9ig (1966; 1975), [A third MBA building in the central part of the mound at Acemnéyik was excavated and sampled by Prof. Allye ‘Oztan last summer, but the samples, sent to us the week before the Bochum meeting in October 2002, have so far turned out to be undatable. Bump 5 (ca. 1832 BC) - Kiiltepe, Wargama Saray! The extra litle blip or bump one sees on the left of Figure 1s from the timbers in the Warsama Sarayt at Kultepe (1999; and earlier dendrochronological comment by Ku- riholm and Newton 1990), the construction date of which is 1832 +4/-7 cal BC. Repair timbers, however, extending down to at least 1771 (but with no bark preserved) show thatthe palace had a life-span of at least 61 years before the burning. For the most recent statement about the Chronology of the Assyrian Colony Period and the length- ening of the Karum II period, see Veenhof 2003, but note that he employs the old dendrochronological dates from the 1996 Nature report Metallurgy and Dendrochronology in the Second Millennium ‘The sites we showed the audience at Bochum in October 2002 were sampled and studied so that we could build a tree-ring chronology. They were not necessarily selected because they had metallurgical potential, although some ‘of them, such as Kaltepe, clearly do. tis one of our con- tinved frustrations that the charcoal emerging from the Karum, either levels Il or Ib, at Kiltepe is usually so bad- ly and thoroughly burned that no dendrochronological dating is possible. We keep visiting the Karum every sum- ‘mer, however, in the hope that suitable charcoal will some day emerge, But now the long master tree-ring chronology exists and ata placement likely not very far from absolute. It should be possible to plug in any site with well-preserved char- coal, including sites from the millennia on either side of the second, Just before the Bochum meeting we received fan e-mail from Walter Gauss in Salzburg to the effect that Inthe Aegina excavations in the 1970s considerable quan- tities of charcoal were discovered in a Kupterschmelzofen ‘and sent to Austria for radiocarbon dating. They were then put on a shelf and forgotten until Dr. Gauss rediscovered them, So he has packed them and sent them to us, and ‘we will See what we can do to date the oven’ 45Peter | Kunlhol, Maryanne W. Newion, Carol B. Griggs & Pamela J. Sulivan| Minor Sites in or near the Second Millennium We use the term “minor” for sites which have few sam- ples, or ones with difficult ring-sequences where not al of the probloms have been worked out. Moreover, i is easy to work with a site from which several hundred samples have been collected. Sites which produce one, or two, oF three samples, it not impossible, are dificult at best. (One site which is not really minor nor a really a problem Js Karahdyik bei Konya where the last existing ring is 1768 +4/-7 cal BC, at least eight years later than AcemhoyOk, We wish only that more charcoal had been saved from this extraordinary site (Alp 1991; 1992), At least three pieces of wood (MAS-4,6,8) from Magat Hoyik date from 1375 +4/-7 cal BC. This is particularly interesting because of the existence of LHIIIA pottery mixed in with the Hitite pots (Ozgug 1978). Aso in central Anatolia we have a date of 1529 (withthe bark present) for Building “C” at Kugakl, but an additional piace recently collected has @ last ring (no bark) at 1523. For an update on Kusakli see the most recent MOOG (Miller Karpe etal, 2003, and Kuninolm & Newton therein). ‘The Hitte site of Ortakby/Sapinuwa is stil giving us prob- lems, and we are holding back on reporting dates until we {determine what is going on there, but the excavator, Prot. ‘Aygil Suel, has now reported major quantities of charcoal collected in 2002 and we will retrieve them in the summer ‘of 2003. Tile Héyok, currently submerged under the Euphrates, has a last preserved ring in the gateway from 1123 +4/-7, cal BC. Although the Tile wood is oak, we are confident ofthe crossdate with the junipers and pines of the Bronze Ageliron Age chronolagy (Kunihoim et al. 1983, but note that these original 1990 calculations do not take into ac- ‘count the revised chronology) ‘Another site in Central Anatolia with longulived oak, Ka- man-Kalenoylk Ild, has a last-preserved-ring in 884 BC +4/-7 (Newton & Kuniholm 2001; annual excavation re- ports summarized in Omura 2001). Out east in Urartu, we have dates for Adilcevaz with its last ring at 776 +4/-7 cal BC (but no bark), Yukari Anzaf at 807 441-7 cal BC, and we have just been promised ‘wood from Karmir Blur next to Yerevan which ought to ccrossdate with both these sites and Ayanis. Down south at Kilisetepe, Nicholas Posigate has a last preserved ring at 1408 +4/-7 cal BC, Another site from this general region, the Kas/Uluburun shipwreck, Is having its cedar sequence wiggle-matched at Heidelberg to confirm or refute our original placement in the 14" century 46 In the west, in the Athenian Agora, are two juniper bed legs excavated in the 1930s in a well that was sealed in the 4" quarter of the 6" century BC (Shear 1940). The last existing ring is 781 +4/-7 cal BC, We know that there is crossdating across the Aegean, but the fit between ‘Athens and Gordion is so extraordinarily good, that one wonders whether Midas in addition to sending his wood- €en throne to Delphi, @ thing well worth looking at accor ding to Herodotos, might have senta wooden bed to Athens, 100, Finally, also in the west, specifically from Shaft Grave V at Mycenae, there is a wooden bowl which was measured years ago under less than optimum circumstances. It was. ‘measured after a long transatlantic airplane fight with the measurements taken off the surface of the bow, and one, would have to go back to the National Museum in Athens In order to remeasure it. It is a single piece of wood, and who knows how fong it had been around when it was: placed in the grave. We have played down this piece for some years as nothing more than a curiosity, but we are happy with a last-preserved ring at 1602 +4/-7 cal BC. ‘The real answer to the Shaft Grave question isin the Na- tional Museum where there isa joist from one ofthe graves. wih an estimated 200 rings, saved by Schliemann and cut off by his carpenter. Polishing and measuring that ‘would be much more satisfactory Acknowledgments ‘The Malcolm and Carolyn Wiener Laboratory for Aegean ‘and Near Eastern Dendrochronology at Cornell Universi ty Is supported by the National Science Foundation, the Malcolm H. Wiener Foundation, and individual Patrons of the Aegean Dendrochronology Project. For fundamental research permissions we thank the ap- propriate governmental and religious authorities in all the ‘countries in which we work, as well as the many excava- tors who not only take time out to explain the intricacies ‘of their sites but who make us welcome at their excava- tion houses year after year. For invaluable assistance in the preparation of this paper we thank Berd Kromer and Sturt ManningDendrochronological Dating in Anatole: The Second Milennium BC Not For al the dates cted above, any change must be consistent. ‘That's to say one cannot have +4 years for one date and ~7 years for another. i=3 years turns Out fo be the fina solution, then thal is =a years across the spoctrum of cited dates trouge hhout the millennium. The graph must move as a whole. Filing Inthe gaps inthe ADIBG transon period and relying on conv ‘elenal dendrochronological dating Instead of radiocarbon wig ‘le-matcning would of course elmnate the nood for this +4/-7 timate. 1 See the March 2003 Aniquity onsine at (htt! aniquiy.ac.ukProjGal/MannngManning him), also. our ‘webiste (hp!/Avaw arts comel.eduldendro) Tor the actual ‘umeers ofthe rathocarbon determinations on which these Calculations are based 2 See ntp:antquty ac uk/PrjGaliDeVies/DeVies. hms 3. At Bochum Prof. JD. Muhly took serous issue with this ‘dentiicaton by the excavators, but whether ti a Kupfer ‘Schmelzofen of not, @ date might stl tll us something ab- ‘out the period fo ihc thie constuction ~ whatever I 8 belongs: Bibliography aps. {S61 Kona arahoyak 1980 Kaz, Kast Sonilan “Gold a akara, 810-98 1002 KonyaKarahoyik 1081 Kaz. Kazi Sonulan rs BW, 901-908. GILINGIROGLU, A. & SALVINI.M, $001: ayans I Ten Yoars' Excavations at Rusahini Eidurs: ‘ai, (Oocumenta Asiana W), Rome: CNR: Istituto per (91 Studi Micene! ed Ego0- Anatole, 377-380. DEVRIES, K., KUNIHOLM, PL, SAMS, G.K. & VOIGT, MM: 2003" New gates for te dSstrucion level at Gordon. Antig- uy 77 no. 296 June 2003) ntp:antiquiy. 2c. kro} GalleeiesiGevees hind HAMMER, CU. 2003) ’Compeling Evidence fora Thera Eropton in 1645 8C Sra Baro Boworonce SCIEN S000 enn. KAOMER, B., MANNING, S.W/, KUNIHOLM, PL, NEWTON, Man SPURK, MLB LEVIN | 2001" gional "CO, gradients in the troposphere: magni- luge, mochanisis. and consequences, Science 234, 2 (December 2001), 2529-2582 KUNIHOLM, PL 1996. The Penistone Aagean:Dandochrongonal Progress as of 1995. Acta Archaeologica 67. 527-835, UMIHOL, PL NEWTON. Mw A677 Year Long Tee-Ring Cronclgy fr the Mie Bronze Age. Studs. in’ Honor of Tahsin. Gzgbe. (Gra: orc Tarts Kurdma Basin), 279-208, 2001 Dendrechronological investigations at Ayanis: Dati the Forrese of usa I: Ausahint Egurrkal. In ® Glingrogia & 6. Savin (ods), supra, 377-380, 2002 Dendroctronological avestigations at Kusakl/Sarssa. Inv A MilerKarpe, ina, 259948, 2008 A revised dendroctrenologial date forthe fortress ot Fueailat Ayana’ sah Esa: Gi pou aM. Sai ec). i eral Age ‘Symposium (n poss KUNHOLM, PL, TARTER, Si, NEWTON, MU. & GRIGGS, Sov Dondrochronoogical nvestgatins at PoreulUuk7a, ‘Turkey 057-1605, Syria 6 44, 370-968, KUNIHOLM, Pi, TARTER, SL & GRIGGS, CB. 402" Ta Hoyok Asjarin) DenrocronciogcalRepor Ins. Summers, ina KUNIHOLM, Pl, KROMER, 8, MANNING, S.W, NEWTON, M., LATIN, CE. & BRUCE, Mi. 1996" “Anatolian re rings and the absolute cronclogy ofthe feastem Mediterranean, 2200-718 BC. Nature 381 (27 ‘Kine 1996) 780-783, MANNING, SW, KROMER, 8, KUNIHOLM, Pl & NEWTON, 260i Anatolian tee-fings and a new chronology for tho east ‘Mersterranean Bronze-lron Ages. Science 294,21 (De ‘ember 2001) 2532-2535. Confirmation of nearabsolue dating of east Master. ranean Bronze-(fon Denarochronology. AMtQuty 77. 2008 fo, 295 (March 2003). "tp/antguiy. ac uk Proj. Galvhlanningimanning hum MOLLER-KARPE, A. KUNIHOLM, PL, NEWTON, MW. & WIL- HELM, G 2002 ‘Untersuchungen in Kugakl: 2001, Mitoiungen dor Deutschen Oront-Geseischatt zu Bern 134, 331-351, NEWTON, MW, & KUNIHOLM, PL 2001" Dendroshronologeal Investigations at Kaman-Kale- hoy Dating Ear wor, Age Leva, Anton Ar chaeological Stuctes Xt OMUAA, s, 200t. Kaman-Kalohohak 10 Anatolian Archaeological Stud- fg 10), Japanese Insitute of Anatotan Arehaeoigy. “Toby. oza0¢, N 1966 ‘Acernhayik Kazilan. Anadolu Xt 1979 Some Contrbutons to Early Anatolian, An from ‘Acempayak," Tur Tarn Kurumu, Bolten XLII, 168, 260-208 oza06.1. 1978” Excavations at Maga! Hoyik and Investigations in its Visit Arava, Tork Tart Kuru 1999 The Palaces and Temples of Kiltepe-KaniyNeya Arcata, Tork Tan Kuru ELON, P. 4992 “Quatre Campagnes & Porsuk de 1986 & 1989. Syria Dax 306-347, REIMER, P: 2001" A’New Twistin the Radiocarbon Tale. Science 204, 21 (December 2001) 2498-2408. SHEAR, TL. 1940" The Campaign of 1999. Hesperia 8, 261-908, STUIVER, M. REIMER, PJ, BARD, E.. BECK, J.W, BURR, G'S. HUGHEN, KA, KROMIER, B., McCORMAG: G.,"van der BLIGHT, Ja SBUAK. M. 1996" “INTGALS® radiocarbon age calibration, 24,000-0 cal BP. Radiocarbon 40, 1041-1086. ‘SUMMERS, GD. 3888" Tile HOyik 4 British Insitute of Archaeology at Ankara. Oxbow (Oxtora, 179-188, \VEENHOF, k 2008" The Old Assyrian List of Year Eponyms from Karum Aanth and Cronotogca intentions Arar Ton Tarin Kuruma, ee Le 47