14C Dates from Tel Rehov: Iron-Age Chronology, Pharaohs, and Hebrew Kings [Science 300, No. 5617 (11 April 2003), 315-318]
Hendrik J. Bruins, Ben-Gurion University of the Negev, Jacob Blaustein Institute for Desert Research, Department Man in the Desert, Sede Boker Campus, 84990, Israel
Johannes van der Plicht, University of Groningen, Centre for Isotope Research, Nijenborgh 4, 9747 AG Groningen, Netherlands
Amihai Mazar, The Hebrew University of Jerusalem, Institute of Archaeology, Mount Scopus, Jerusalem 91904, Israel
27 January 2003; accepted 11 March 2003
Stratified radiocarbon dates provide an independent chronological link between archaeological layers and historical data. The invasion by Pharaoh Shoshenq I (Shishak) is a key historical synchronism, ~925 B.C.E., mentioned in both Egyptian inscriptions and the Hebrew Bible. The list of places raided by Shoshenq, mentioned at Karnak (Egypt), includes Rehov (Israel). The site yielded a consistent series of radiocarbon dates from the 12th to 9th century B.C.E. Our results (i) suggest a revised Iron-Age chronology; (ii) date an archaeological stratum to Shoshenq's campaign; (iii) indicate the similarity of "Solomonic" and "Omride" pottery; and (iv) provide correlation with Greece and Cyprus.
The difficulty in providing secure linkages between Egyptian history and archaeology in the Levant, particularly in relation to the Bronze and Iron Ages, as well as in relation to biblical data, has long been frustrating. Chronology in these periods is largely based on historical Egyptian data, indirectly linked to archaeological strata through complex association based on artifacts. Various interpretations relating to age assignment and historical association may be possible. The Iron Age chronology in the Near East and Eastern Mediterranean Region has become a hotly debated issue (1-6). Certain archaeological assemblages in key sites in Israel-for example, Hazor, Megiddo, and Gezer (Fig. 1) were traditionally assigned to the time of King Solomon (second half of the 10th century B.C.E.). The association of such pottery assemblages with biblical data relating to the United Monarchy of David and Solomon was considered reasonably secure, functioning as a ceramic benchmark to assess the archaeological picture of this period, as well as the subsequent Divided Monarchies ofIsrael and Judah. However, the discovery of similar pottery in the destruction layer of Jezreel, a royal citadel of the Omride Dynasty, which ruled over the northern Kingdom of Israel in the 9th century B.C.E. (885 to 843 B.C.E.) (77), posed a problem. This discovery led to the proposal of a low chronology for the Iron Age (2, 5), in which the boundary between Iron I and Iron IIA is lowered by about 80 years, from ~1000 B.C.E. to ~920 B.C.E. The implication is that archaeological layers traditionally associated with the United Monarchy of David and Solomon "have become" too young. Thus an archaeological controversy arose, in which chronology is pivotal.
Radiocarbon dating was not seriously used for Iron Age sites of the Near East, because the method initially lacked the required precision in historical years (8). As high-precision calibration curves based on dendrochronology (9-13) became available, the need for an independent chronology in Near Eastern archaeology based on 14C dating was advocated (14). Advances were made (15, 16) and radiocarbon dates have been obtained for Iron Age sites (17, 18), but inconsistencies and wide age ranges gave ambiguous results.
Here we report a stratified series of highquality radiocarbon dates from Tel Rehov, in northern Israel (Fig. 1; table S1), ranging from the 12th to the 9th century B.C.E. We have used the most accurate procedures afforded by 14C dating: single-year organic samples taken only from primary contexts, high-precision dating, multiple measurements, conventional and accelerator mass spectrometry (AMS) techniques (19). The calibration precision in historical years was greatly enhanced by stratified archaeological wiggle matching (20, 21), because successive layers cannot have the same position on the calibration curve but must follow each other in time.
Tel Rehov is located in the Beth-Shean/ Jordan Valley, some 6 km west of the River Jordan (Fig. 1). This area constituted a major junction in the Iron Age, providing east-west and north-south routes of international importance. The tell has a size of 10 hectares, and is thus one of the largest Iron Age sites in Israel (22). The place name Rehov appears in various Egyptian New Kingdom texts. Most important, it occurs on the list of cities con quered by Pharaoh Shoshenq I, the biblical Shishak (I Kings 14:25-26, II Chronicles 12:3-4). His Asian campaign was recorded on the southern wall of the temple of Amun at Karnak in Upper Egypt, where Rehov appears after the term "The Valley" (probably referring to the Beth-Shean! Jordan Valley) and before the city name Beth-Shean (23). This sequence fits the local geography well (Fig. 1).
The excavations reveal that Tel Rehov was a thriving city during the Iron Age I and IIA cultural periods (22). The city was destroyed and rebuilt several times, which left a series of occupation strata. Three of the strata (numbered VI to IV) belong to the Iron Age IIA, containing the same kinds of artifacts found in other key sites mentioned above in relation to the Iron Age chronological controversy. Earlier layers (Strata D-6 to D-4 in excavation area D) contain ceramic assemblages typical for the Iron Age 1. We present our dating results (tables SI and S2; Figs. 2 and 3), going from young to old in the stratigraphic sequence (24).
At the end of the Iron Age IIA, half the site was abandoned after a severe destruction of Stratum IV. A large building of Stratum IV uncovered in Area C contained rich finds, including many pottery vessels typical for Iron Age IIA, cultic objects, and a rare Greek Middle Geometric vessel. Charred cereal grains found on the floor in thick destruction debris of this building were dated by AMS, because three measurements were made on subdivided samples. The dating results were consistent, within 1σ of each other. The weighted average date [2755 ± 25 years before the 14C present (yr B.P.)] gives a 1σ calibrated age range of 918 to 892 yr B.C.E. with 25.4% relative probability and another age range of 880 to 836 yr B.C.E. with 42.8% relative probability (Fig. 2). The calibration curve descends steeply and regularly during the second half of the 10th century B.C.E. and the first two decades of the 9th century (Fig. 3). Then the calibration curve goes up around 875 B.C.E. to form a small plateau that lasts until 845 B.C.E. Hence, there are two principal options for the calibrated date of Stratum IV. The period 880 to 836 B.C.E. is most likely in probability terms, but 918 to 892 B.C.E. is also possible. The invasion of the Aramean ruler Ben Hadad I during the time of King Baasha of Israel (902 to 886 B.C.E.; I Kings 15:20) is a possible candidate. But other events following the end of the Omride Dynasty seem more plausible for the destruction of Stratum IV and the abandonment of the lower city. The Jehu revolt (843 B.C.E.), the Assyrian invasion of Shalmaneser III (841 B.C.E.), or the Aramean invasions of Israel during the time of Hazael (between 840 and 830 B.C.E.) all fit the radiocarbon dating results.
Stratum V at Tel Rehov stratigraphically predates Stratum IV, although the pottery assemblages from these two strata are almost indistinguishable. Greek Proto-geometric pottery found in this level is of great importance for establishing correlation with Greek chronology. The destruction of this city is found in several parts of the tell, and Stratum V yielded the largest amounts of charred grain. Three consistent 14C dates from Area B (locus 4218) gave a weighted average of 2786 ± 25 yr B.P. Three different loci of Area C also yielded consistent dates and weighted averages of 2771 ± 8 yr B.P., 2788 ± 14 yr B.P., and 2776 ± 9 yr B.P., respectively. The 1σ and 2σ calibrated age ranges (table S1; Fig. 2) of the various loci of Stratum V all give a distinct highest relative probability for the 10th century B.C.E., and particularly the period 935 to 898 B.C.E. These 14C dating results match well with dates suggested by Egyptologists for the reign of Pharaoh Shoshenq I, ~945 to 924 B.C.E. or slightly younger (25), and for the year of his invasion into Israel, tentatively suggested as 925 B.C.E. (25) or 918 B.C.E. (7) in correlation with biblical texts. Whatever the merits of the latter options, all possibilities fit well with our radiocarbon dates. Placing our dating results of Stratum V on the calibration curve (Fig. 3), also with respect to Strata IV and VI, leaves no reasonable alternative but the period 940 to 900 calendar years B.C.E. Therefore, we attribute the destruction of Stratum V at Tel Rehov to Shoshenq I, as there seems to be no other historical candidate that would fit the available radiocarbon time window.
We did a computer simulation using the OxCal program (26) to evaluate which radiocarbon dates (in years B.P., with a standard deviation of 20 years) and calibrated age ranges one would obtain if 925 B.C.E. were the date for the Shoshenq invasion of Canaan. A total of 60 simulations gave individual radiocarbon dates similar to the ones we obtained for Stratum V, as well as older and younger dates. The average of the 925 B.C.E. simulations is 2795 ± 20 yr B.P., which is close to our results from Stratum V. Indeed, the graphical presentation of the simulated calibrated date for 925 B.C.E. is nearly the same as our dates for Stratum V (Fig. 2). Our 14C results would also allow for a slightly younger date of Shoshenq's campaign within the 10th century B.C.E., because the entire period 940 to 900 B.C.E. is possible in terms of radiocarbon dating.
Four radiocarbon dates were obtained for Stratum VI. One conventional date obtained by gas counter on seeds gave a result of 2761 ± 14 yr B.P. (GrN-27366). Two AMS dates of a mixture of seeds and fine charcoal from the same basket gave older dates, 2805 ± 35 yr B.P. (GrA-21054) and 2800 ± 50 yr B.P. (GrA-21182); a third measurement yielded a young date of 2755 ± 35 yr B.P. (GrA-21043). There is one wiggle in the calibration curve from 950 to 980 B.C.E., which seems to fit the 14C dating results for Stratum VI, in relation to the archaeological stratigraphy. The pottery assemblage of this city belongs to an early phase of the Iron Age IIA, but it differs only slightly from that of Strata V and IV. It seems that Stratum VI should be dated to the first half of the 10th century B.C.E., and perhaps closer to its middle part, as indicated by 14C dating.
Area D at Tel Rehov is a stratigraphic trench excavated in the western slope of the lower mound. It shows a series of eight wellstratified strata, ranging from the end of the Late Bronze Age (Stratum D-8) until the Iron Age IIA (Phase D-l and D-2). Phase D-2 consists of refuse deposits from nearby Area C Strata V and VI. The only date we obtained for D-2 (GrN-26112, 2805 ± 15 yr B.P., measured on olive pits), 10th century B.C.E, fits well chronologically with refuse from either Stratum V or VI.
Phase D-3 yielded five consistent individual radiocarbon dates from three different loci, all measured on charred olive pits found in refuse or storage pits, situated below the above-mentioned D-2 refuse. The weighted average is 2831 ± 18 yr B.P., which favors a calibrated age range for the end of the lith century B.C.E. and the first half of the 10th century B.C.E. Placing the result in stratigraphic order on the calibration curve would fit the range of 1010 to 980 B.C.E. (Fig. 3). The only alternative possible period, 960 to 940 B.C.E., does not fit in terms of stratigraphic wiggle matching with respect to D-4 and Stratum VI. Moreover, the Late Iron Age I ceramic assemblage of D-3 is generally accepted to predate the Iron Age II-A pottery of Stratum VI.
Phases D-4a and D-4b constitute, respectively, a street surface and parts of two houses. Two consistent dates on olive pits from Phase D-4a, measured by conventional gas counter (GrN-26121, 2890 ± 30 yr B.P.) and AMS (GrA-18825, 2870 ± 50 yr B.P.), give a weighted average of 2885 ± 26 yr B.P. The calibrated date gives a wide possible age range for the 12th and lith century B.C.E. because the calibration curve in the 11 th century is rather level (plateau), whereas many wiggles occur between 3000 and 2900 yr B.P. for the period 1250 B.C.E. to 1130 B.C.E. However, putting our 14C dating results in stratigraphic sequence on the calibration curve favors the period 1050 to 1010 B.C.E. The ceramic assemblage of phase D-4a is typical Iron Age I-B pottery. Three AMS dates of seeds trom locus 1845 of Phase D-4b are consistent and give a weighted average of 2924 ± 22 yr B.P. Though there are four to five possible calibrated range options in the 2IT period 1254 to 1020 calendar years B.C.E, the period 1090 to 1050 B.C.E. seems to fit best in stratigraphic sequence on the calibration curve (Fig. 3).
Phase D-6 (upper) consists of occupation debris above a floor (locus 2836) in which charred olive pits were found. Two consistent dates, both measured by conventional gas counter (GrN-26118, 2920 ± 30 yr B.P.) and AMS (GrA-18826, 2950 ± 50 yr B.P.), give a weighted average of 2928 ± 26 yr B.P. The lσ calibrated range lists five possible periods in between 1209 to 1050 B.C.E. Considering the stratigraphic sequence and the ceramic assemblage, the period 1130 to 1090 B.C.E. on the calibration curve seems most likely (Fig. 3). The pottery is typical for the second half of the 12th century, probably slightly after the end of New Kingdom Egyptian presence in parts of Canaan, which occurred at some time in the period 1150 to 1135 B.C.E., during the reigns of Pharaohs Ramesses IV to VI (23, 25).
In conclusion, the radiocarbon results, in relation to archaeological, historical, and biblical data, lead us to propose a modified traditional chronology for the Iron Age in the Levant (table S2). The modification is that the Iron Age IIA cultural period includes both the 10th and much of the 9th century B.C.E. (~980 to 835 B.C.E). There is only one known historical candidate that fits the destruction date of Tel Rehov Stratum V, 940 to 900 B.C.E., based on 12 high-quality 14C dates: the invasion of Ph araoh Shoshenq I.
Our research negates an important argument of the low chronology theory, namely, that Iron Age IIA ceramic assemblages should be confined exclusively to the 9th century B.C.E. The 14C dating results imply that it is difficult to distinguish between "Solomonic" and "Omride" pottery. The site of Ta'anach (27), about 8 km southeast of Megiddo (Fig. 1), is also mentioned on the Karnak list of places destroyed by Shoshenq. Period II-B pottery at Ta'anach, assigned to 960 to 918 B.C.E. (27) and to the 9th century in the low chronology (28), is identical to that found in Tel Rehov Stratum V. Period II-B ended in a fierce destruction, which can be related to Shoshenq's campaign in view of our results.
Because Shishak (Shoshenq I) is mentioned as a contemporary of Solomon in biblical texts, we find it plausible to retain the linkage of specified archaeological assemblages (Rehov Stratum V, Ta'anach II-B, Hazor X, Megiddo VB, and perhaps also V A-IVB, etc.) to the United Hebrew Monarchy. Our results also have implications for the chronology of Cyprus and Greece because imported pottery trom both countries was found in Tel Rehov Strata V and IV. It appears that the traditional chronology of Greece can be maintained, but for Cyprus, older dates seem appropriate for some pottery groups (29, 30).
Materials and Methods
The dated organic material, charred cereal grains and olive pits, with one exception of charcoal, was derived from recent excavations (1997-2000) conducted by Mazar (S1). Short-lived botanical remains were discovered in each of the excavated strata at Tel Rehov in primary context, including some refuse pits. This provided a unique opportunity to obrain a series of stratified high-quality radiocarbon dates for the entire sequence, suited for stratigraphic archaeological wiggle matching (S2, S3) on the calibration curve, in order to advance the establishment of an independent chronology in Near Eastern archaeology based on 14C dating (S4, S5), S6).
Besides the obvious advantages of short-lived seeds, sometimes these little grains may slip out of their original stratiographic context, due to biological activity or processes. The consistency of the results gives reassurance concerning the position of most charred seeds as "in situ" regarding their respective stratigraphic contexts.
All radiocarbon dates were measured at the Radiocarbon Laboratory of the Center for Isotope Research, University of Groningen (The Netherlands). The charred organic samples were treated by the acid/alkali/acid (AAA) method (S7). The purified organic matter of each sample was converted into CO2. The 14C content of the CO2 of the larger samples (gram size C) is measured conventionally in gas counters. The CO2 from small samples (mg size C) was transformed into solid carbon for measurement by Accelerator Mass Spectrometry (AMS) (S8).
The conventional method is preferred where possible, because measurements with the highest precision can be achieved: a standard deviation (1σ) of 10-15 years for 25 grams of sample. The (1σ) precision of the AMS measurements is 30-50 year. Triplicate / duplicate measurements of subdivided samples were used to increase accuracy and precision (S6). Moreover, the average value of consistent multiple dates yields a high-quality date with a smaller standard deviation. This is important for the calibration from radiocarbon years into calendar years (S9, S10).
References and Notes
4. S. Gitin, A. Malar, E. Stern, Eds. Mediterranean Peoples in Transition Thirteenth to Early Tenth Centuries BCE (Israel Exploration Society, Jerusalem, 1998). (back)
7. M. Miller, J. Hayes, A History of Ancient Israel and Judah (Westminster, Philadelphia, 1986). (back)
11. 1993 Calibration issue, M. Stuiver, A. Long, R. S. Kra, J. M. Devine, Eds., Radiocarbon 35 (1993), 35-65. (back)
12. INTCAL 98: Calibration Issue, M. Stuiver, J. van der Plicht, Eds., Radiocarbon 40 (1998). (back)
15. 'Near East Chronology: Archaeology and Environment', H. J. Bruins, I. Carmi, E. Boaretto, Eds., Radiocarbon 43 (2001). (back)
23. K. A. Kitchen, The Third Intermediate Period in Egypt (1100 BC) (Aris & Phillips, Warminster, ed. 2, 1986). (back)
24. Owing to length restrictions, we are unable to present a more detailed description of each sample in its stratigraphic context, ceramic assemblages, or a discussion of published radiocarbon dates. (back)
25. K. A. Kitchen, in The Synchronisation of Civilisations in the Eastern Mediterranean in the Second Millennium B.C. (Österreichischen Akademie der Wissenschaften, Vienna, 2000), 39-52. (back)
26. C. Bronk Ramsey, Radiocarbon 37, 425 (1995). We used OxCal version 3.5 (2000) with a resolution of 4 and without rounded-off ranges. (back)
31. We are grateful to J. Camp for supporting the excavations at Tel Rehov. Our thanks to H.-J. Streurman, A. T. Aerts-Bijma, and S. Wijma for carefully preparing and measuring the radiocarbon samples. (back)
S7. W.G. Mook, and H.T. Waterbolk, Handbook for Archaeologists. No. 3. Radiocarbon Dating. (European Science Foundation, Strasbourg, 1985). (back)
S8. J, van der Plicht, S. Wijma, A.T. Aerts, M.H. Pertuisot, H.A.J. Meijer, Nuclear Instruments and Methods B172 (2000), 58-65. (back)
S9. M. Stuiver, J. van der Plicht, Eds., INTCAL 98: Calibration Issue, Radiocarbon 40(3) (1998). (back)
S10. C. Bronk Ramsey, Radiocarbon 37 (2), 425-430 (1995). We used OxCal version 3.5 (2000) with resolution = 4 and without rounded off ranges. (back)
Table S1. Radiocarbon dates of Tel Rehov in relation to the stratigraphy and archaeological context. GrN = (Groningen) conventional measurement by gas counter; GrA = (Groningen) AMS measurement. Calibrated dates are presented only for the respective average dates. (back)
Table S2. The radiocarbon dating results, following calibration and stratified archaeological wiggle matching, in relation to historical, biblical, and archaeological data. Our revised traditional chronology for the Iron Age of the Levant is compared with the low chronology. All dates are in years BCE. (back)