The CANEW Project

Data compiled by Laurens THISSEN and Damien BISCHOFF
Last updated: 12/02/2006

All calibrated data quoted at 1sigma. Calibration curve used is IntCal04 (Reimer et al. 2004). Calibration program used OxCal v3.10 (Bronk Ramsey 1995, 2001, 2005).

Zawi Chemi-Shanidar: In expressing the Zawi age on the chart we used the Rubin/Alexander date as the more reliable information, being from the lab itself.

Abu Hureyra 1: The Abu Hureyra 1 range is composed as follows: Sum phase 1 (exclusive of OxA–883 (too old) and OxA–882 (too young), so remains OxA–172 and the R_Combined date OxA–468+387+469+470 = 10,977±76 BP) is 11,100–10,880 cal BC (68.2%). Sum phase 2 (exclusive of the 5 dates failing at 5% in the X2-test, and excl. of OxA–6996 (too young)), remains a total of 6 dates (the R_Combined set counting as 1): 11,210–10,700 cal BC (68.2%). A single date for phase 2-3: 10930–10820. Sum phase 3 (excl. of OxA–8718 (too old) and of the 3 R_Combined dates failing at 5%), remains 3 dates: 10,930–10,610 (54,6%); 10,550–10,440 (13.6%). The total possible range within which the Abu Hureyra 1 occupation most likely might have occurred: 11,100–10,440 cal BC (12 dates used in total).

Mureybet: Sum IA (Natufien):10,300–9400 (6 dates); sum II (PPNA) 10,800–10,400 (8.5%), 10,200–9300 (59.7%) (5 dates); sum IIIA (PPNA): 9700–9100 (55.8%), 9050–8850 (12.4%) (4 dates); sum IIIB (PPNA): 9300–8600 (7 dates); sum IVA (PPNB): 9200–8800 (13.7%), 8700–7900 (54.5%) (3 dates); sum IVB (PPNB): 8700–7950 cal BC (2 dates). Hence: Natufian Mureybet and earlier PPNA (level II) overlap fully date-wise. The total range may be set at 10,300–7950 cal BC.

Çayönü (A): Treated here as a separate site, Çayönü Round Buildings Subphase r1 relates both as to architecture and material culture to Hallan Çemi I–III, M'lefaat and Nemrik. Despite having errors larger than ±200, the two r1 dates are in line with the general 14C ranges of these settlements.

Hallan Çemi: The range, based on the SUM of all dates, minus the three which stem from low quantity carbon (BETA–47252, BETA–47253 and BETA–56102), is constellated as follows: 10,800–10,400 (12.7%) and 10,100–9200 (55.5%), where the relative probability increases strongly between 9800–9200 cal BC, this being the most likely age range for Hallan Çemi.

Qermez Dere: The Qermez Dere data are hard to assess. Note the mutual exclusiveness of the two pre-6 data, though on short-lived material and stemming from the same context; note also the indication of the level 6 sample being stratigraphically identical to the level 5 sample, but again their 1s ranges are mutually exclusive. Date-wise but crosscutting level boundaries there are two clusters: 1) OxA–3752 (level pre-6) and OxA–3756 (level 5), and 2) OxA–3755, 3757, 3754. Of these mutually exclusive data events the earlier one sums at 10,050–9450 cal BC, the younger one at 9260–8840 cal BC. OxA–3753 seems much too old and is not used in the graph.

Göbekli Tepe: The SUM value of the two archaeological dates is 9130–8650 cal BC. Combining them would yield 9120–8760 cal BC. It is this latter value which is used in the chart, since the samples are likely to stem from a single event.

M'lefaat: Apart from the huge error margins, the four dates above 11,000 are even less acceptable when considering their stratigraphic location, i.e. from lower phase house 8 fill, and from upper phase house 3 fill. Ignoring them we note that the two upper phase Gd-dates are older than the oldest lower phase date, making the Gd-data set as a whole weak. Remain two OxA-dates from the lower phase and two from the upper phase. Both being from the same contexts and on identical sample material, their posterior probability distributions may be combined. Lower phase 9650–9200 but with 60.7% prob. at 9460–9200 cal BC (agr. 94.2%); upper phase 9300–8940 but with 63.2% at 9300–9130 cal BC (agr. 88.9%). On this basis we have set the age range within which occupation at M'lefaat may have most likely ocurred as between 9500–9100 cal BC in the chart.

Jerf el Ahmar: The dates without stratigraphic context have not been used in our chart, although overlapping fully with the remaining four dates. The latter sum at 9500–8550 cal BC.

Dja'de: Dates with margin >200 not used. Ly–6166 and Ly–2614 seem too young, not used because low amount of charcoal and lack of collagene resp.; Ly–5821 too old, not included in sum. Individually summed, the age ranges per phase are as follows: Phase I: 9310–8830 (2 dates): 9310–9000 (59.0%), 8920–8830 (9.2%); Phases I–II, II (4 dates): 9150–9000 (8.6%), 8850–8550 (59.6%); Phase III (13 dates): 8540–8510 (6.6%), 8490–8290 (61.6%). Accordingly, the total range for Djade Phases I–III would be 9310–8290 cal BC. The overall SUM total differs considerably given the weight of the Phase III dates, and tends to ignore the early dates, yielding a Dja'de I–III (19 dates used) range of 8640–8270 cal BC, which we will follow on the chart with the reservation that occupation of the site may have started two centuries earlier than that.

Nevali Çori: Summing 19 of the Nevali Çori dates would give a range of 8650–7950 cal BC. The Hd–16784-768 date (9882±224 BP) has not been included, as falling widely out of the general range of dates. As many as nine dates show remarkable overlap, summing at 8550–8290 cal BC. These dates are distributed over all individual levels (Level I: OxA–8303 and 8235; Level I/II: HD–16781-16783; Level IIIA: OxA–8302; Levels 'III–V': KIA–14756, 14762, 14760). Taken together, these dates signal a major occupational event at Nevali Çori in the third quarter of the 9th millennium cal BC.

Çayönü (B): Given a gap of at least 600 calendar years existing between the earliest round buildings (subphase r1) and the present sequence, we treat both units separately. Without knowing the exact provenances of the Çayönü dates and their attribution to specific building levels, every attempt at putting dates on the individual subphases as most recently defined by A. Özdogan (1999), must remain tentative. Especially the last three subphases are insufficiently covered by 14C, and the single date for the Large Room Subphase in the 7th millennium would contradict that phase's position within the PPNC. We give here the SUM ranges for the individual subphases, on which the graph in the chart is loosely based.

A substantial part of the 1sigma SUM ranges for Round Buildings (r3–4), Basal Pits, Grill Buildings–late grills and the Channeled Buildings Subphase overlap roughly between 86/8500–83/8200 cal BC. All later phases, and notably the immediately ensuing Cobble-Paved Buildings Subphase are exclusive of this earlier cluster at the 1sigma level.

Cafer: We have omitted Ly–4436 as probably too old, as well as Ly–3772 and 3773 since contradicting the sequence (too young). The eight remaining dates sum at 8300–7450 cal BC.

Akarçay Tepe: Since only one date is available per level, summing is useless. The total 14C range runs from 7900–6100 cal BC, where it must be noted that all dates are mutually exclusive. Phase VI is as yet undated.

Gritille: The graph shows the summed value of the dates, taking account of the R_Combined results, exclusive of the two dates failing at the X2 test (GrN–15246, GrN–15250), and exclusive of the two dates with errors >200 (GrN–15254, Beta–13215). The remaining nine dates sum at 8000–6800 cal BC (66.6%) and 6700—6600 cal BC(1.6%).

Tell Assouad: MC–607 is not consistent with the general range and therefore not used in the chart. The remaining two dates contradict the sequence. A date somewhere in the 2nd quarter of the 8th mill. seems nonetheless most acceptable – c. 7800–7400 cal BC.

Abu Hureyra 2: In assessing the SUM value of the Abu Hureyra dates OxA–880 and BM–1720R have been left out in the analysis as being much too early for the associated contexts; BM–1122 and BM–1120 have been omitted as taken from two or more different levels (sic), and dates with errors >200 (BM–1723R and OxA–6995) are not included either. The total sum over the remaining 30 dates is at 7750–7050 cal BC.

Halula: Eight dates with errors >200 have not been used in assessing the SUM range (UBAR–396, Beta–58931, UBAR–384, UBAR–394, Beta–70359, UBAR–290, Beta–58926, UBAR–395). Of the remaining ones Beta–61488 (9520±180 BP) seems too old. The remaining seven dates sum at 7800–6400 cal BC.

Tell Sabi Abyad II: Sum is at 7650–6750 cal BC, but all three dates are mutually exclusive and stemming from different levels.

Hayaz: Two mutually exclusive dates at 1sigma, where the younger date (GrN–12512) due to its large error has not been used in assessment of the date range in our chart, and we tend to the 7500–7200 cal BC range suggested by the older date to cover Hayaz's second layer as well.

Bouqras: The total of dates allows for the distinction of four age groups: 1) 7500–7300 cal BC (GrN–8263, 8262, 13104, 13103, 13102, 13080); 2) 7200–7000 (GrN–8261, 8258, 4818, 10592, 13100); 3) 7000–6600 (GrN–8260, 8259, 8264, 4819, 4820, 13090, 10591, 13099); 4) 6400–63/6200 cal BC (GrN–13082, 13081, 13079, 10589) These crosscut all levels and phases of Bouqras and hardly ever are consistent with the stratigraphic sequence with one notable exception related to age group 1: the basal deposits of the reset 1975 section, as well as the lower levels of the deepsounding (9–8) do yield the oldest data of the site, thus providing a secure terminus ante quem for the approximate beginning of Bouqras.

Shimshara: Summing the dates exclusive of K–981, which is evidently too old, yields 7050–6050 cal BC. The two dates from levels 13 and 11 overlap, and are mutually exclusive of the level 9 date.

Mezraa/Teleilat: AA–49102 is much too old, and not used in the sum calculation. Given the nature of the samples and their provenance we used the Combine function for the individual loci, combining the posterior probability distributions of the dates. Three dates from locus 38 combine as 7050–6820 cal BC (agr. 116.8%); three dates from locus 62 result in 7040–6820 (106.5%); 2 dates from locus 100 give 6650–6530 cal BC (115.3%). Level IIB is thus much younger than level IIIB2 which is securely datable in the first centuries of the 7th mill. cal BC. The basal level IV above virgin soil is as yet undated.

El-Kowm 2-Caracol: The sum range is calculated without Ly–4438, 2521 and 4439 due to their excessive error margins, where the remaining four dates yield a value of 7100–6000 cal BC.

Tell Damishliyya: The oldest date and the two later dates are mutually exclusive at 1sigma, although deriving from the same context. Given the presence of older deposits at the site, UtC–1094/–1095/–1096 might represent pre-5 levels; its value is therefore retained on the chart suggesting this earlier occupation.

Telul eth-Thalathat: TK–199b and TK–199a were not used in the chart because of the large error margins. They also are not consistent with the other level XVI date, nor with the level XV dates. Total sum 6700–6100 cal BC.

Tell Sabi Abyad I: Assessment of the Sabi Abyad I dates, based on a combination of the SUM parameter and R_Combine (in the case of sample deposition at the same moment or within a short span of events), allows Operations III (and IV) to be ranging between 6430–6230 cal BC; Operation V between 6210–6090 cal BC, and Operation I between 6030–5800 cal BC, where the Operation II dates are less consistent, but show an overlap with I, which is confirmed culturally. Below the Operations III and IV deposits deeper levels have already been excavated, but dates are not yet available (Peter Akkermans, pers. comm. 14 Dec. 2005).

Since the R_Combination of the two Magzaliya dates fails at 5% in a X2-test, this site has not been added to the chart. Also Jarmo is not occurring on the chart, as no clear statements can be made on the 14C evidence, and no summing is attempted because of contradictions of dates and sequence, the insecurity of provenance and level assignation, and the poor overall quality of dates. The Jarmo dates do suggest an occupation somewhere between 7000–5000 cal BC, but we have refrained from adding this to our chart. Finally, also Qdeir is not included, because its one radiocarbon date has too large an error as to be of use.