Milestone Achievement in Our Research: Quantitative Analysis of First Temple Period Pottery

Last Thursday, we reached a significant milestone in our research and publication process. At the annual New Studies in the Archaeology of Jerusalem and its Region conference, organized by the Israel Antiquities Authority, the Hebrew University, and Tel Aviv University, we presented key findings from the quantitative analysis of pottery from the mid-late First Temple Period (Iron Age IIB-C). Our study focuses on comparing these artifacts with finds from other Jerusalem sites, particularly the Ophel area.

For the first time, we employed data gathered from sampling soil from various locations around the Old City, processed, sifted, and cataloged in the same rigorous way we use for the Temple Mount soil. These samples provide essential control groups for statistically analyzing the frequencies of finds from the Temple Mount.

We recently conducted this comparison because our project differs from standard excavations, where finds can typically be processed and published incrementally by each excavated area. In our work, each artifact is part of an enormous puzzle: sorting and classifying one find influences the interpretation of others. Only in recent years have we reached a level of completeness in cataloging pottery and coins that has enabled us to conduct high-resolution typological studies and finally perform quantitative analysis.

Looking back at our control groups from other parts of Jerusalem, we now understand that sampling additional locations could have revealed even more data patterns. This type of rigorous sampling may seem commonplace to readers with backgrounds in exact sciences, but in archaeology, particularly in our region, quantitative comparisons between typological artifact groups using statistical tests like Chi-square are rare. Studies that aim to examine the broader picture usually gather data from published excavation reports across multiple sites, an approach that is often biased due to variations in excavation methods and the volume of material recovered at each site.

Our analysis suggests that artifacts from fills have unique potential for statistical research. Fills cover large spatial areas and are more dispersed, unlike undisturbed contexts, which generally produce concentrated, restorable pottery from a single structure or short time frame. Artifacts from fills can thus offer a broader perspective on regional activity over extended periods, rather than narrowly representing individual sites.

In our First Temple Period pottery analysis from Temple Mount soil, we found many similarities to the Ophel excavation, located just south of the Temple Mount. This pattern is unsurprising, as both sites were part of the city’s Royal Compound. In the First Temple Period, the Temple Mount—often referred to in the Bible as Mount Zion—had different boundaries from those of the Haram al-Sharif compound we see today, which reflect Herod’s later building and expansion projects. According to biblical sources, there was no extensive sacred precinct surrounding the Temple at that time. While the Bible mentions courtyards near the Temple, these were likely functional spaces for priests, housing elements such as the altar of burnt offering and the Sea of bronze. Nearby were administrative and public buildings such as the king’s palace, the House of the Forest of Lebanon, and the Hall of Judgment.

In the Ophel excavations, especially in the recent work led by the late Eilat Mazar, fortifications were uncovered, including walls, towers, a gatehouse, and public buildings. This suggests that the Ophel likely marked the southern boundary of the Royal Compound, with the Temple at its northern end. For this reason, the First Temple Period finds from both the Temple Mount soil and Ophel area likely reflect public structures and perhaps even pilgrimage-related activity, including visitors from regions beyond Judah.

These findings, though ambitious, are grounded in observed trends, confirmed through statistical methods, and represent our initial interpretations. As we continue our studies, sample further locations and additional Iron Age IIB-C  finds from excavation reports from Jerusalem, we expect to uncover further insights and potentially new interpretations. Regardless, this study holds considerable value, especially for its methodological innovation.

The complete results of our analysis will be published in the Iron Age IIB-C pottery chapter of the final Sifting Project report.

For more details, see:

• Below: Presentation text (Hebrew with AI English translation).
• Published article (Hebrew with English abstract)

---

Good afternoon to everyone.

Thanks to the organizers of the conference…

The presentation is on behalf of myself and Gabriel Barkai, my partner in the Temple Mount Sifting Project, whose health unfortunately prevents him from attending today. We wish him a full recovery. It is also on behalf of Hillel Richman, who authored the typological report on Iron Age ceramics, and who also cannot be here today due to a family event.

The research we are presenting today is not a typical publication of the Temple Mount Sifting Project but rather represents a milestone in analyzing the abundant findings accumulated over the project’s twenty years. This is reflected in the development of data analysis methods and in our understanding of the soil we are sifting.

Background

We assume that most of the audience here is familiar with the background of the Temple Mount Sifting Project, so there is no need to elaborate. However, we see fit to give a brief overview of the project’s research methodology, as it is innovative and was developed in response to the question of how to extract the maximum data from artifacts that were not discovered in situ.

When we transported the soil from the main dump in the Kidron Valley, we divided it into different areas. The purpose of this division was to distinguish between areas that may have been disturbed by prior soil dumping and those that clearly represented undisturbed material from the Temple Mount. Additionally, we conducted manual excavations to identify the boundaries of the Temple Mount’s soil dump. The soil from the Temple Mount is characterized by a grayish hue (due to the high ash content) and is saturated with significant quantities of glass tesserae and glazed wall tiles from the Dome of the Rock’s exterior, as well as with plastic prayer mats.

The soil originates from the southeastern part of the Temple Mount and was removed during illegal construction work. Although we were confident that the soil indeed came from the Temple Mount, the question remained about the nature of the soil itself—namely, whether it represents activities on the Temple Mount or whether it was brought to the Mount at some point in the past, containing archaeological material from elsewhere.

As research progressed and data accumulated, it became clear that the ground level in the eastern area of the Temple Mount was significantly lower in the medieval period than today (approximately 8 to 12 meters). From the Mamluk period onward, this area accumulated soil and stone dumps resulting from construction, renovations, and clearing of debris on the Temple Mount. Thus, these are not purposefully placed fills but rather multi-layered accumulations of soil and stones originating from various areas on the Mount.

The question about the origin of pre-Herodian finds remained. Could it be that during the Second Temple period, soil containing archaeological material from another site was brought to the Temple Mount? Our hypothesis is that the soil brought at that time likely came from the eastern slopes of the Temple Mount—a close, accessible, and uninhabited area. This suggestion was also previously made by A. de Groot and H. Geva.

However, following the discovery of an Iron Age garbage pit at the foot of the Temple Mount’s eastern slopes, it is possible that this area contained additional trash pits from that period, likely originating from the Temple Mount itself. Artifacts from the First Temple period in the soil from the Temple Mount may, at least in part, derive from these trash pits. This refuse originally likely came from the Temple Mount and was returned to it with soil fills during the Second Temple period.

Research Approach to Artifacts from Soil Fills

Artifacts found in soil fills and soil accumulations can have great research value when viewed as a statistical sample representing a broad area, especially when collected through wet sifting, which yields 100% of the finds in the soil. For this reason, we deemed it necessary to sample soil dumps from additional locations to create statistical control groups for data analysis from the Temple Mount. In these samples, soil from eight sites around the Old City was sifted using the same methods as those for the Temple Mount soil, and cataloging and sorting were done in the same way.

In our comparative study, the aim was to examine any significant differences between the finds from the Temple Mount and those from the samples and to compare these against quantitative data from other Jerusalem excavations. However, it is challenging to find publications with such detailed data. Most ceramic publications from Jerusalem excavations lack quantitative information, and if such information exists, it is typically not sampled based on a uniform criterion.

An exceptional publication on this topic recently appeared in Ariel Winderbaum’s PhD. dissertation, which dealt with Iron Age pottery from Ophel excavations led by the late Eilat Mazar.

The analysis of Iron Age ceramics has been conducted by Hillel Richman for the past decade. To avoid statistical bias, the classification focused on rim sherds rather than diagnostic body sherds. A quantitative analysis appendix was added to the typological report, examining unusual phenomena in the distribution of pottery types, both in comparison to samples from other Jerusalem sites and within the excavation areas in the Temple Mount soil dumps.

The statistical analysis was performed using SAS software and utilized advanced data mining tools.

Analysis of the Iron Age IIB-C Pottery from Temple Mount Soil

Approximately 5% of the ceramic finds from the Temple Mount soil date to Iron Age IIb-c, with an additional 1% dating to Iron Age IIa. The distribution of Iron Age IIb-c ceramic types from the Temple Mount in relation to samples from Jerusalem showed a similar general distribution of vessel types between the two groups. However, a series of differences emerged, some of which also appeared in the Ophel excavation data. Here is a selection of these differences:

First, we present the general distribution of the types of vessels (Class). It appears that there is a strong similarity between the distributions in the samples; however, two significant statistical differences emerged. The prevalence of jars on the Temple Mount is significantly higher compared to control samples from Jerusalem, while the percentage of lamps in the city-wide samples is significantly higher than on the Temple Mount. This could be because the Temple Mount area mainly contained public buildings, which included storage areas with many jars. On the other hand, lamps are more common in residential areas.

Similarly, in the Ophel data, jars from the Iron Age IIb–c horizons appear more frequently than in earlier periods. Moreover, the percentage of lamps from the late Iron Age IIb and Iron Age IIc is notably lower than in the other samples. It is reasonable to assume that public buildings also used lamps, but their usage was likely more common in residential homes. Regarding the Temple and its immediate vicinity, biblical descriptions suggest there was little nighttime activity. In addition, the sanctuary itself was illuminated by golden lamps (Menoras).

The main quantitative analysis focused on comparing the distribution of vessel types across different classes. Here, too, we often found similar distributions between the Temple Mount soil and the samples from Jerusalem, though there were also many differences, which were confirmed statistically.

For example, we found considerable variation between the distributions of jug types from the Temple Mount and the control samples. Types JG 1 and JG 7 are significantly more common on the Temple Mount compared to the control groups samples, while other types of jugs are more prevalent in the control groups. Type JG 1 is the second most common jug in the Temple Mount soil, and its prevalence there is five times greater than in the other samples. This jug is known as a cooking jug, with soot remnants observed on some examples. It was more common in the early Iron Age IIa and continued to appear in certain sites in Judah and Jerusalem during Iron Age IIb. Its prevalence is also relatively high in the Ophel, where it constitutes about 31% of all jugs from the Iron Age IIb–c horizons. We tried to examine the prevalence of this type in other excavations using ceramic plates and found that the cooking jug’s prevalence was significantly lower in those sites compared to the Temple Mount and the Ophel.

Based on this data, it seems that for some reason, cooking jugs were particularly prevalent on the Temple Mount and in the Ophel. The specific cooking method used with these jugs is still unclear, but research by Ben-Shlomo and others suggests that cooking with them was done without direct contact with fire. The jug was likely placed near a heat source, allowing for prolonged simmering of liquid dishes.

Another phenomenon observed in the jug data was the significantly lower prevalence of red-slipped jugs on the Temple Mount compared to the control samples. Ceramic plates from Ophel excavations show that red-slipped jugs account for about 12% of all jugs, compared to the much higher percentages reported in other Jerusalem excavation reports.

The types of cooking pots also revealed a particularly interesting trend. Only the most common type, KR1, appeared in the samples, while three additional types were found on the Temple Mount. This does not imply that other types of pots were unused in other parts of Jerusalem; rather, their relative frequency is low. Therefore, a sample like this, which includes only 19 fragments, may not trace their prevalence.

These additional types are characterized by smaller dimensions than KR1 and higher-quality finishing, especially in their red-slipped, burnished exterior. These pots are known from other sites in Judah, as well as coastal and Philistine sites (and from other areas in the country, including the north). The data raises the question of whether this reflects a wider ceramic diversity on the Temple Mount and a preference for smaller pots, possibly indicating the site’s uniqueness as a focal pilgrimage point for all of Judah (and possibly the entire country).

An intriguing phenomenon was also discovered in the cooking pot data. The main difference was in the CP1 type, which was very common on the Temple Mount but did not appear in the control samples, despite being found in Jerusalem excavations. This cooking pot has a rim with a decorated ridge creating a stepped look. One example of this type from the Temple Mount was painted with a thick white paste. A similar white paste was observed on another rim of a jug from the Temple Mount and on several sherds from the Ophel. Excavators there suggested that this phenomenon might be related to ritual use of the vessels due to the resemblance to Judean figurines, which were also often painted white. To our knowledge, this phenomenon is unique and, aside from examples from the Ophel and Temple Mount, is unknown elsewhere.

The CP1 type is very common in the Shephelah, southern Judah, and Samaria, and as mentioned, was also found in Jerusalem. This raises the question of why it did not appear in our Jerusalem samples. Additionally, it was challenging to determine its prevalence in Jerusalem excavations due to publications lacking quantitative data. However, the Ophel data presents a complex picture. In Windebaum’s publication, this type does not appear at all, while in the publication by Mazar and Morgan on ceramics from Area A2009, three out of seven cooking pots were of this type. It appears we need to await the full publication of Iron Age IIb–c horizons to better estimate this type’s prevalence in the Ophel area.

The similarity between Iron Age IIb-c finds from Temple Mount Sifting Project and Ophel excavations also extended to other types of finds. For example, we found an unusually high percentage of clay sealings with fabric impressions on their backs in the Temple Mount and in the Ophel. In Temple Mount clay sealings, they make up about 25%, while in the Ophel, they account for 40%, compared to just 6% in all Jerusalem excavations. In an article we published in 2021, we suggested that this type of clay sealings was primarily used in treasury administration.

A comprehensive study by A. Greener on figurines from Temple Mount Sifting Project revealed that figurine types from sites around the Temple Mount differ from other Jerusalem and Judean sites in that they lack non-Judean-style figurines (which are typically found in small quantities at Judean sites), such as relief figurines, hollow figurines, and so on. Figurine assemblages from Temple Mount soil, all Ophel excavation publications, and a refuse pit discovered on the eastern slopes of the Temple Mount were included. Greener, following others, suggested explaining this phenomenon by a preference for figurines with Judean traditions and an ideological resistance to figurines with foreign influences.

Discussion and Summary

The similarity we demonstrated between Iron Age finds from Temple Mount Sifting Project and Ophel excavations helps answer the question about the provenance of the finds—that is, whether these artifacts indeed represent activity on the Temple Mount or if they came from another site, brought there with soil fill brought in from outside during the Second Temple period.

It is essential to remember that today’s Temple Mount boundaries, created as part of Herod’s expansion, do not reflect the First Temple period’s boundaries. According to the biblical sources, there was no defined sacred precinct around the Temple as there was in the Second Temple period. The Temple was part of a complex of government buildings on the southeast hill of the city, known today as the “Royal Compound.” This area was fortified; findings from Ophel excavations, including a wall, gatehouse, fortifications, and public buildings, indicate that this area formed the southern boundary of the Royal Compound of the city, which was fortified.

Therefore, it is not surprising that shared characteristics exist between finds from Temple Mount soil and Ophel excavations since both were part of a single administrative complex in the First Temple period. However, we lack archaeological data to determine exactly where the northern and western boundaries of this wall, enclosing the Royal Compound, were located and whether it also served as the city wall or as a separate internal fortification, beyond which additional residential neighborhoods enclosed by the city wall may have existed.

In any case, it seems that the Royal Compound northern boundary did not extend beyond the Bethesda Valley  route, which runs northwest-southeast across the Temple Mount. In the later stages of Iron Age II, the city expanded to the west and perhaps also to the north. These areas contained residential neighborhoods, some of which may initially have been homes built outside the walls. Therefore, it is unsurprising that Temple Mount soil contains artifacts typical of residential areas, though they appear less frequently than in samples representing primarily the western part of the city.

Here, we presented an innovative methodological approach demonstrating that artifacts from fill deposits and debris accumulation layers have significant value when analyzed using quantitative methods, providing a wide spatial overview. Soil fills also allow proper sampling sampling, enabling data analysis using statistical tools.

In other words, data of finds which were recovered out of  stratified context such as fills, which do not represent a specific structure but rather a wider area and contains a greater amount of finds, provides a comprehensive and representative picture of the material culture and activities in the region. Additionally, fill deposits, especially refuse features, include a broader chronological range than sealed layers, which generally represent only the final stage of a particular period.

We presented here, for the first time, a statistical analysis of artifact data accumulated over many years of Temple Mount Sifting Project. These results are initial findings of the study, indicating the great potential in the research method employed.

Processing the finds from this project involves tedious work that requires considerable patience. However, it is expected to yield valuable results that will further illuminate the Temple Mount’s history and material culture.

Acknowledgments: Many partners contributed to this research, and it is impossible to mention them all. We offer our deepest gratitude to those who helped with funding and research, especially the hundreds of thousands of visitors who participated in the sifting.