Compositional Analysis of Cultic Clay Objects from the Iron Age Southern Levant

Abstract

Compositional analysis conducted on pottery and other ceramic items can shed light on their place of production and in certain cases, on technological aspects of the production sequence. The methods used, petrography and chemical analysis, can also be employed on cultic terracotta such as figurines, cult stands, models, or other clay objects. Several studies of such analyses of items from various periods in the Southern Levant have been published, mostly from temple contexts. This paper focuses particularly on two groups of items: clay models from the favissa at Yavneh and pillar figurines and other (mostly horse) figurines from Jerusalem and Tell en-Nasbeh in Iron Age Judah. These two groups are both roughly dated to the time span between the 9th and 7th centuries BCE. While the former group is of objects representing a temple context in Philistia, the latter is likely related to a domestic cult in Judah. The analysis of these objects is also examined against the background of a robust compositional analysis of regular pottery from the sites. The compositional analysis can indicate whether these objects were locally produced or imported from various regions (thus possibly brought by pilgrims), as well as whether they were “mass-produced” in a single workshop. The results can shed light on aspects of religious and cultic conducts in these occasions as well as compare domestic and temple-related cultic behaviors.

1. Introduction

The archaeological evidence for the religion and cult of the ancient Southern Levant includes artifacts, often made of clay, that were used for rituals and/or reflect in their iconography local beliefs and religion. Compositional analysis conducted on these ancient clay artifacts can shed light on their place of production and in certain cases, also on technological aspects of the production sequence. The main methods used are petrography and chemical analysis, and these can also be employed on cultic items made of clay such as figurines, libation vessels, stands, and models. The compositional analysis can improve our understanding of ancient cults and religions in various ways. The results can indicate whether these objects were locally produced, possibly by a temple workshop, or imported from various regions, and thus possibly brought by pilgrims. In relation to domestic cultic artifacts they may indicate whether they were “mass-produced” in a single, specialized workshop or produced in a regular pottery local workshop. If they turn out to be imported, this may indicate the movement of people, as these are often considered as private items. Thus, the results can shed light on aspects of religious and cultic conducts in these occasions as well as compare domestic and temple related cultic behaviors.

Clay cultic items from various periods in the Southern Levant have been analyzed on several occasions, though not commonly. Most publications deal with artifacts from temple contexts including those from the Iron Age.

This paper focuses particularly on two groups of items from Iron Age Israel: 1. cultic (shrine) models from the favissa at Yavneh (Ben-Shlomo and Gorzalczany 2010); 2. pillar figurines and other (mostly horse) figurines from domestic contexts at Jerusalem (Ben-Shlomo and Darby 2014) and Tell en-Nasbeh (Ben-Shlomo and McCormick 2021), both in Iron Age Judah (these are also examined against the background of a robust compositional analysis of regular pottery from the sites). The two groups are both roughly dated to the time span between the 9th and 7th centuries BCE, representing the period of ethnic states in the Iron Age Southern Levant. While the former group is likely related to a temple context in Philistia, the latter represents a domestic cult in Judah.

2. Materials and Methods

The most popular method for the compositional analysis of clay objects, aiming to investigate their place or production (provenance) and technological sequence, is a thin-section petrographic analysis (henceforth, petrography). Samples are obtained by standard thin-sectioning, with a several millimeter-sized sample from the artifact (naturally, for some of these, the sample size had to be quite small), mounted and thinned to about 0.03 mm (30 microns; see, e.g., Quinn 2013). The method is based on the refraction of polarized transmitted light through the various minerals and rock fragments in the clay (raw material of artifact), which become transparent in a thin section. The characterization of the mineralogical composition of the clay through the polarized light microscope can point to its geological origin on the one hand, and to the treatment of the clay and addition of temper by the potter on the other.

The inspection of the thin sections aims to characterize their components, including matrix (fine component under 4 microns) and coarser component (non-plastic inclusions). Matrix characteristics may include optical activity, spacing between particles, and voids. Non-plastic inclusions characteristics may include mineralogical or rock identification, size and shape distribution (sorting), frequency (percentage from slide area), size ranges, and shape of particles. Generally, the samples within a given study are allocated to “petrographic groups” according to the characteristics of the matrix and non-plastic inclusions identifications, Here, however, the petrographic groups (or “petrofabrics”) were defined according to the soil type the clay source was related to (according to maps and descriptions of local soils in the Southern Levant or modern Israel, in most cases, as in Dan et al. 1972; Sneh et al. 1998; Goren 1996; Goren et al. 2004; Singer 2007).

Another method used for finding the provenance of pottery and clay objects is a bulk chemical analysis. In our case, most studies conducted use Neutron Activation Analysis (henceforth, INAA). This method is based on the chemical fingerprinting of the clay source, in particular using the abundance of a large group of trace elements in the clay paste. For INAA, powdered samples from the artifact weighing about 100 mg are irradiated in a nuclear reactor. The elemental abundance is obtained by the interpretation of the Gamma-ray spectrum emitted by the irradiated sample. The provenance is determined by comparing the elemental composition of the samples, using multivariate statistics, to a given data bank containing reference materials (such as samples from kilns or natural clays analyzed) (see, e.g., Mommsen et al. 1988).

3. The Case Studies: Results

3.1. Temple Cult

During 2002, a large favissa located within a pit near Tel Yavneh, some 20 km south of Tel Aviv, was excavated in a salvage excavation directed by Raz Kletter (Kletter et al. 2010). The pit, which was damaged by some modern gardening work, included hundreds of terracotta objects interpreted as house (shrine) models or cultic stands (Figure 1), and other cultic vessels (mostly “pans”, chalices, and other pottery vessels, totaling about 3500 in number). The date of the assemblage is the late Iron IIA to early Iron IIB (9th or early 8th centuries BCE, Kletter et al. 2010, pp. 205–6; Panitz-Cohen 2010, pp. 130–31), and it apparently belonged to a temple from the nearby site (the “Temple Hill”). The iconography of the objects includes depictions of humans (mostly female, Figure 1(3,4)) and animals (mostly bovines but also lions, Figure 1(1–3)), as well as vegetative and architectural depictions (Figure 1(2,4)). The iconography is generally similar to the Canaanite tradition cultic representations; however, Philistine iconographic elements, resembling Iron I figurines (such as Figure 1(3) see Ziffer and Kletter 2007, pp. 19, 28–29; Kletter et al. 2010, pp. 86–91; Kletter et al. 2015) also appear. The assemblage includes at least 119 figurative architectural (shrine) models and a few dozen other figurative objects, in addition to thousands of “pans” or “shovels”, chalices, and bowls, together with other types of mundane pottery vessels. The Yavneh assemblage is a an unusually rich cultic assemblage in Iron Age Israel and quite unique. The site is physically located within the Philistine territory (e.g., 2 Chr 26:6), and it is chronologically dated to a stage (the Iron Age IIA) when Aegean attributes are less evident in the material culture of the Philistine, yet a distinct culture is maintained (see, e.g., Maeir et al. 2013).

Petrographic Results

The Yavneh assemblage analyzed by petrography included 133 samples altogether (Ben-Shlomo and Gorzalczany 2010). Of these, the following types were sampled: 75 shrine models, 7 fire pans or ‘shovels’, 16 chalices, 4 various cultic objects (figurines, etc.), and 31 pottery vessels, mostly bowls and juglets. The majority of the samples (102 samples) can be considered a component of the cultic assemblage, while the other items appear to be mundane vessels but were probably also part of the paraphernalia of the temple and may have also been used in cultic-related practices but could have been used for other purposes as well. The petrographic analysis indicated a high homogeneity in the fabric of the ceramic objects from Yavneh, (Figure 2 and Figure 3). About 116–119 of the 133 samples (87%) indicated a similar source of clay source, and only minor variability appeared (Figure 2, Group 1 in Figure 3; “hamra” group, Ben-Shlomo and Gorzalczany 2010, Figure 9.2; Group 1, with Sub-Groups 1a–1e). Only fourteen samples indicated different sources (see Figure 3).

The fabric which dominated the Yavneh assemblage, denoted as the “ḥamra” group (Figure 1; Ben-Shlomo and Gorzalczany 2010: Group 1, Figure 9.2), was non-calcareous and relatively compact. Non-plastics included mainly quartz, with rounded to sub-rounded sand-sized particles (sized 0.15–0.4 mm, “beach sand”) very few heavier minerals (residual from Nile sediment), and a lower number of calcareous inclusions. This ḥamra soil represented in the clay can come from outcrops which are abundantly located in the vicinity of the site (the more northern Mediterranean coast of Israel has a higher amount of calcareous sand) (Dan et al. 1972, p. 44; Dan et al. 1975). The source of this clay was presumed to be the central coastal plains of Israel or local to the site of Yavneh. Ḥamra soil was used on several occasions as a major raw material by ancient potters (e.g., Goren et al. 2004, pp. 292–94, Pl. XII:EA294,EA296). There is no evidence for intentionally adding quartz sand (since its distribution seems natural) to the clay, and this clay was probably not highly levigated. Yet, certain levigation treatments of the clay, at least in some of the samples, could possibly have altered the natural occurrence of the quartz sand in some manner.

Minor differences in this large group appeared, however; these were defined accordingly as sub-groups (also Ben-Shlomo and Gorzalczany 2010, Table 9.1, Groups 1A, 1B, 1C, 1D and 1E), yet, that group was still quite homogenous. From 75 stands, 37 belonged to the regular high-fired ḥamra group (Figure 2a, Group 1A), 23 samples belonged to a fabric which was basically similar but was probably fired at a lower temperature (Figure 2b, Ben-Shlomo and Gorzalczany 2010, Group 1B; see also Figure 2c,d for other variants), and only 15 belonged to other fabrics, including 6 to 8 samples made of loess-type soil from the southern coastal plains (Figure 3, Group 2; Ben-Shlomo and Gorzalczany 2010, p. 146, Group 2, provenanced to the southern coastal plains) and 2 to 4 examples were imported from outside the Southern Levant (Ben-Shlomo and Gorzalczany 2010, p. 146, Groups 4 and 7, 1 or 2 came from the Northern Levant coast of Phoenicia, and one possibly from Cyprus). It should be noted that the pottery vessels analyzed from Yavneh (mainly bowls and chalices) were also almost all made of the same ḥamra group clay (Figure 3; Ben-Shlomo and Gorzalczany 2010, p. 147, Table 9.2).

The petrographic results thus indicated that the Yavneh assemblage, especially the models (at least 80% of them were analyzed), were almost all locally made and were probably made in one workshop using a relatively similar clay recipe also used for other pottery vessels (see Figure 3, Group 1 ḥamra). Due to their high thickness of sandy clay, the fired models were still quite brittle and did not appear to be highly fired ceramic objects. Although the different models showed a rich and diverse iconographic world, it seemed that the stands and other cultic vessels were “mass-produced” in a single workshop in the vicinity of the site and were not brought by pilgrims from a distance. These objects were probably used for some ceremony in the temple (though, they rarely have soot marks) and then discarded after a short time; this may explain their brittle nature, as they were possibly not intended for continuous use (see below).

3.2. Domestic Cult

3.2.1. Figurines from Jerusalem

Iron Age clay figurines from Jerusalem were analyzed by petrography (Ben-Shlomo and Darby 2014; Ben-Shlomo 2019, pp. 243–46, Figures 7.49–7.51). These included 103 figurines from the City of David (

Table 1. Figurines analyzed from Jerusalem.

Type	City of David (Y. Shiloh)	City of David (E. Mazar)	Ein Hamevaser	Total
Judean pillar figurines	40	16	3	59
Horse figurines (of these, horse and rider)	13 (1)	20 (3)	0	33 (4)
Other zoomorphic figurines	6	5	0	11
Total	59	41	3	103

, both from Y. Shiloh and E. Mazar excavations) and 10 Iron Age IIA items from the Ophel (Ben-Shlomo 2015). Most figurines were either anthropomorphic female figurines (Figure 4; Judean Pillar Figurines = JPF, and other types) or zoomorphic (mostly depicting horses, Figure 4(10–12)).

Judean pillar figurines depict rather schematic standing women, with only the head, neck, hands and breasts detailed. The head is either made by a mold (Figure 4(1–5)), more naturalistically depicted with details such as hair and jewelry, or hand-made, showing a “pinched” bird-like face (Figure 4(6–8)). The body, made separately, is cylindrical and usually hollow. These figurines are one of the distinct components of the cultural material of Iron Age Judah and are extremely frequent in Jerusalem (many hundreds of examples were published), appearing in lower quantities as one withdraws from the city (e.g., Kletter 1996; Darby 2014). These figurines may be a part of the Yahwistic religion of Judah in the 8th–6th c. BCE. Several researchers suggested that the JPFs were associated with the domestic cult of Asherah, understood as the consort of Yahweh (Olyan 1988; Kletter 1996, p. 180).

Another group, which is even more common in Judah and Jerusalem, is horse and horse and rider figurines (Figure 4(10–12)). These figurines are also schematically modeled, showing a cylindrical head with few details and flat, bed-like body with four conical legs attached. A small group includes a rider attached to the body (also schematically depicted). Iron Age II zoomorphic figurines have been studied to a lesser extent (see, e.g., Holland 1995; Amr 1980; Press 2012, pp. 183–86, 252–60). The animal figurines are usually also interpreted as cultic, but usually, their function and meaning seem less clear (e.g., the large group coming from Cave I in Jerusalem; Holland 1995). Recently, the horse and rider figurines were also interpreted as a substitute for the depiction of the god Yahweh since several of these were found in the Iron Age IIA temple at Moẓa near Jerusalem (Garfinkel 2020).

3.2.2. Petrographic Analysis

The composition of the figurines was compared to a large group of Iron Age pottery vessels analyzed from Jerusalem (426 vessels, mostly tableware, cooking pots and storage jars, Ben-Shlomo 2019).

The majority of the figurines represented a generally similar clay type (Figure 5 and Figure 6; Ben-Shlomo and Darby 2014, Group 1, Figure 3; 67–68 figurines). This was a calcareous, fossil-rich clay, likely derived from local rendzina soils developing on soft calcareous soils (chalk) (Figure 5(1–5); similar to Ben-Shlomo 2019, Group 7, Figure 7.50(1–5)). This generic soil type is common in various regions in Israel such as the central hills, the Shephelah, and the Galilee (Levantine Ceramic Project [LCP] Petrofabric Nos. 7 and 24 or 71 https://www.levantineceramics.org/petrofabrics/7-israel-palestinian-authority-calcareous-foraminiferous-rendzina-soil-on-upper-cretaceous-formations, accessed on 19 May 2025). The clay used, however, showed variability in the composition and could be subdivided into at least five sub-groups (Figure 5(1–5); Ben-Shlomo and Darby 2014: Groups 1a–1e) with somewhat different texture and quantity of non-plastics: high dolomite and calcite inclusions (e.g., Figure 5(4)); a finer porous matrix and many small argillaceous inclusions (e.g., Figure 5(2); see Whitbread 1986); a darker matrix (higher firing temperature; as Figure 5(1)); or a fabric very rich in microfossils (as Figure 5(5)); these sub-groups were also not very homogenous. Clay mixing may also be indicated by the appearance of clay pellets or “balls” (Figure 5(2)) variable in size and amount. The composition of the figurines may represent the variability in clay procedures used in the same workshop by the same or different potters, or several different workshops, although generally, the clay is assumed to come from roughly similar sources (see below).

Second in frequency in the raw materials used for the figurines was a non-calcareous clay with a reddish matrix, fewer microfossils, and a higher quantity of quartz, mostly silty, as well as calcareous inclusions (Figure 5(7)). The clay was related to terra rossa soil (Ben-Shlomo and Darby 2014, Group 3, Figure 4(2), a total of 21 figurines), the most common local clay type used for tableware and cooking pots in Jerusalem during the Iron Age (Ben-Shlomo 2019, Group 1). While this group was not very homogeneous, it represented a different workshop, or clay recipe, than that of the rendzina clay. In addition, eight of the City of David figurines were made of dolomitic Moẓa marl or Moẓa clay with dolomite sand (Figure 5(6); Figure 6; Ben-Shlomo and Darby 2014, Group 2, Figure 4(1)), a local clay which was commonly used for storage jars and other vessels in Iron Age Jerusalem (Ben-Shlomo 2019).

Five or six figurines analyzed represented sources beyond Jerusalem or the central hills; these were made of clay related to loess soils common on the southern coastal plain (Philistia) (Figure 5(8); Ben-Shlomo and Darby 2014, Group 5, Figure 4(3)). Note that pinched heads and molded heads of figurines were made of similar clays. The zoomorphic (horse) figurines were even slightly more variable in their clay than female figurines, but, generally, they had a similar composition (Ben-Shlomo and Darby 2014, Figure 5).

A smaller earlier petrographic study of 15 figurines from the City of David suggested most were made of a similar clay (Goren et al. 1996, p. 88), while an INAA analysis of 18 figurines (Yellin 1996) yielded two chemical groups varying in Ca and Fe.

The clay recipes used for figurines in Jerusalem may indicate they were not produced by potters making other pottery vessels. However, there is no ethnographic or ancient textual evidence for this phenomenon in the Near East (e.g., London 1989, p. 78; Duistermaat 2008, p. 354). Another option is that figurines were less “marketable” than pottery vessels, and therefore, poor quality clay was used in their production (also less effort was given in preparing and homogenizing the clay, and they may have been fired at a lower temperature). Another option is that certain clays were chosen due to tradition or due to ritual or symbolic reasons (e.g., Sterner 1992, pp. 171–79, in Cameroon). See more discussion below.

3.2.3. Tell En-Naṣbeh

The site of Tell en-Naṣbeh lies today withing the modern city of Ramallah and is about 12 km northwest of Jerusalem. The region has various formations of the Judea group from the Cretaceous era (Shachnai 2000), as well as terra rossa soils, derived from hard calcareous rocks (Dan et al. 1975).

Tell en-Naṣbeh is second to Jerusalem regarding the quantity of JPFs retrieved (at least 186 in number, Kletter 1996; Darby 2014, pp. 244–45). The site was excavated during the 1930’s by Badè (e.g., McCown 1947). Due to the excavation methods, many figurines did not come from very clear stratified contexts, yet, they are easily typologically identified. Fifteen JPFs, one male figurine fragment, and five bed models were analyzed by petrography (Ben-Shlomo and McCormick 2021; these were part of a group of 186 JPFs analyzed typologically by McCormick; see Ben-Shlomo and McCormick 2021, Appendix 1). Samples were chosen to represent the collection. The specimens came from cisterns, rooms, and a dump. Mostly heads were sampled, as they featured more typological variability.

Fourteen of the fifteen figurines’ samples from Tell en-Nasbeh and five “bed models” were made of the same type of clay most of the Jerusalem figurines were made of: rendzina-type clay (Figure 7, Group 10; Ben-Shlomo and McCormick 2021, Figure 7). Rendzina-type soils occur only about 10–25 km south–southeast of Tell en-Naṣbeh in the Jerusalem region (see below, Figure 8) or in farther areas (see, e.g., Ben-Shlomo and Darby 2014, pp. 187–90). Furthermore, of the 200 Iron Age pottery vessels and other ceramic objects analyzed from Tell en-Naṣbeh, hardly any were made of the rendzina clay (Figure 7; Ben-Shlomo 2021, Group 10; 200 samples of pottery vessels and terracotta, mostly Iron Age, were analyzed by petrography; several were also analyzed by INAA). Thus, rendzina-type soils are likely not local to the site of Tell en-Naṣbeh, and the objects made of this clay were likely imported from the Jerusalem region. Possibly, some of these figurines may have also been brought by refugees coming from Jerusalem during the final Iron Age and Babylonian periods by people of Jerusalem exiled by Nebuchadnezzar from 597 BCE onwards.

One of the figurines and one of the “bed models” were made of the more commonly used terra rossa-type clay (Figure 7, Group 1).

4. Discussion

The petrographic analysis of a large group of cultic objects, mostly architectural shrine models, from the temple-related favissa at Yavneh, indicated a local, rather homogenous production. This may indicate that these cultic items may have been produced in a workshop also related to the temple. Other examples of Bronze Age temple workshops (for pottery and metals) come, for example, from Hazor and Nahariya (e.g., Susnow and Yahalom-Mack 2023). In our case, the origin of the objects cannot shed light on the identity or origin of the people using the temple even if they were pilgrims. The shrine models show variability in the iconography and compositions depicted: possibly the models represent the temple at Yavneh or a temple in general. They are less likely to represent domestic structures due to the animals (bulls, lions) and figures (musicians, nude women, trees) standing in their entrances (Figure 1(3,4); Ziffer and Kletter 2007). These may have been purchased or ordered by believers visiting the temple and used, for example, for more private rituals or ceremonies, or possibly were part of the temple’s paraphernalia for more public ceremonies. In any case, they were discarded after use.

Several other studies discuss compositional analyses of cultic assemblages related to temples in the Levant, including Iron Age temples or shrines. The late Iron Age cultic pottery assemblages of the shrines at Ḥorvat Qitmit and Ḥaẓeva (also interpreted as a favissa assemblage, Ben-Arieh 2011) were analyzed by petrography and INAA. The INAA results from Qitmit indicated a similarity between the clays used for cultic items and regular pottery (Gunneweg and Mommsen 1995); a similar result was obtained for the Ḥaẓeva objects according to both petrography and INAA (Gunneweg et al. 1991; Cohen-Weinberger 2011). Anthropomorphic stands or models and cultic vessels from the Ḥaẓeva favissa were all (except one) made of a local and homogenous clay (Cohen-Weinberger 2011, p. 188). Several cultic vessels were analyzed by INAA from the Iron Age I temple of Tell Qasile (Yellin and Gunneweg 1985); the analysis indicated these were made of local clays. As no petrographic analysis was conducted on most cultic vessels from Tell Qasile, their other technological characteristics remain unclear.

Another assemblage from a temple that was studied by petrography includes 53 pottery vessels from the Late Bronze Age temple at Lachish (this temple on the tell was excavated by the Hebrew University; see Weissbein et al. 2019). Most of the vessels were locally made from loess-type clay. Yet, as the assemblage hardly contained cultic terracotta and has not been published yet, more details cannot be given here.

Objects from much earlier Chalcolithic cultic centers, such as the En Gedi shrine, were analyzed by petrography (Goren 1995). These revealed a different picture of more non-local objects, yet, the social structure of the Judean Desert Chalcolithic people, likely to be nomadic and tribal, is clearly very different from Iron Age, more urban societies represented by temples built in cities.

Studies from the Aegean region are also often focused on temple assemblages, such as the Cypriote-style figurines from the Late Iron Age Aegean (Karageorghis et al. 2009), and the Late Minoan/Sub-Geometric Syme Viannou temple (Nodarou and Rathossi 2008). These examples illustrate mostly locally made cultic objects, but also significant variability in clay types and treatment.

Thus, so far, we see that Iron Age temple sites of the Southern Levant used cultic vessels that were produced locally, often made of rather standard clay also used for regular pottery vessels. The phenomenon of vessels brought from distant regions to the temple, possibly by pilgrims, is rarely attested. It may be difficult to assess according to compositional analyses whether the selection of clays, forming of cultic items, and/or their plastic decoration, rich with iconography, were conducted by different potters or artisans in the same region or workshop, other than the regular pottery vessels’ potters. Possibly, “regular” potters who worked in a workshop providing the temple also specialized in this artistic work. It is also often difficult to prove by compositional analysis alone that temples had their own attached pottery workshops. In certain cases, such as at Middle Bronze Age Nahariya and Tell el Hayyat, there is archaeological evidence for this (Susnow and Yahalom-Mack 2023, pp. 200, 209). There is so far no evidence for special rituals related to the selection of clay or production of objects (as cited in some Mesopotamian texts, e.g., Abusch 2002, and see below).

Fewer compositional analyses were conducted on cultic artifacts from domestic contexts, such as figurines. Previously analyzed samples were only sporadically (and often of relatively rare types) and occasionally published as part of excavation reports (for example, Goren et al. 1996. Thus, the case of Jerusalem figurines is an excellent opportunity for such a wider-scale study and for a comparison between production modes of more official temple cultic objects and those used in domestic cult (see

Table 2. Comparison of analysis of cultic objects from Yavneh and Jerusalem.

Aspect	Yavneh	Jerusalem
Samples analyzed	107	103
Types	Architectural shrine models, pans	Female pillar figurines, horse figurines
Iconography	Rich iconography with many themes; relatively naturalistic	Relatively standard, few themes, schematic depictions
Provenance	Locally produced 95%+ Few from distant sources	Locally produced 90%+ No evidence of distant sources
Homogeneity (clay type)	Very high (ḥamra over 90%)	High (rendzina 70%)
Homogeneity (inclusions)	High	Low
Regional geological variability	Moderate	High
Tempering	Quartz, homogeneous	Calcareous, variable
Firing	Mostly high	Variable, usually relatively low
Comparison to non-cultic pottery	Same production location and technique	Different production: same region, different clays
Interpretation: production mode	Single workshop (temple production?) high production control and standardization	Several workshops (household production?) or multiple potters

for a comparison between Yavneh and Jerusalem). It is not surprising that the domestic cult objects usually found in family households were largely locally made, as the results described above show. Having figurines imported from distant locations in these contexts would probably be interpreted as evidence for immigrants.

According to the Judean female (JPF) and horse figurines’ distribution, it is likely that the Jerusalem region was a center of production of those, and this is so far also corroborated by the compositional analysis; for example, most of the Tell en-Naṣbeh figurines (see above) and three figurines at Mevasseret were near Jerusalem (Darby 2011, p. 310).

However, was there a specialized figurine workshop in Jerusalem related to any form of administrative or religious center? If the JPFs (as well as the horse figurines?) were a normative component of the Judean religion during the late Iron Age, as has been suggested (see, e.g., Kletter 1996; Hadley 2000; Byrne 2004; Dever 2005), the figurines, although used domestically, could have been produced by the royal administration in central production centers.1 Byrne (2004) even associated the JPF with a “national” propaganda encouraging fertility, which would explain their domestic use. Petrographic results seem to indicate a variability in clay treatment and low standardization and thus may rule out that the production of these figurines was in one workshop and production line (as might be expected from Byrne’s views). The fact that Jerusalem may have been the center of the Yahwist, possibly aniconistic, religion may also influence the production mode of the Jerusalem figurines, and being less relevant for the main temple cult activities in Jerusalem, the figurines were not centrally produced. According to the finds from the excavations in Jerusalem and its vicinity, figurines were intensively and continuously used, employed throughout many households in Judah, as more than a single example was found per household. The large quantities of figurines may represent multiple uses on frequent occasions of single figurines, or rarer-use sets of many figurines in a certain house (for communal events?). Again, the compositional results of the figurines’ analysis indicate that the former option is more probable than the latter, due to the low standardization of the clay.

Somewhat similar patterns were revealed in compositional studies carried outside the Levant, such as the Körös Neolithic culture in northeastern Hungary (Kreiter and Szakmány 2011) or the Aztec Empire (Otis Charlton and Charlton 2010, INAA). These indicate that domestic cultic artifacts were made of similar clays to the regular pottery vessels (also Kreiter and Szakmány 2011, p. 116). However, they were actually more standard in their clay than pottery vessels; see also (Gheorgiu 2010) on the standardization in production techniques of Chalcolithic figurines from southeast Europe, (Day et al. 2006) on the analysis of objects from Kavousi, Crete, and (Fillieres et al. 1983) on Athenian Agora figurines.

Nevertheless, the variability in the clay used for the figurines in Jerusalem and Judah should be examined in light of the variability in the natural clay sources in the Jerusalem region, and the clays used for the production of common pottery vessels in the same period. Jerusalem lies on a geographic and climatic border zone between Mediterranean soils and climate in the west, and semi-arid climate and different soils in the east (Figure 8; see, e.g., Sneh and Avni 2011; Ben-Shlomo and Ackerman 2019). This is reflected in the variability in soils and geological formations that can be used for clay collection. Three main options are available in the radius of 10 km from the City of David (Figure 8): Moẓa clay or marl in the west/northwest (Moẓa and Aminadav Formations), terra rossa soils in the west (Weradim and Netzer Formations), and calcareous rendzina soils (Menuha Formation) in the south and east (see, Ben-Shlomo and Ackerman 2019, Figures 3.4 and 3.9; Ben-Shlomo 2019, Figures 7.36–7.38). A large-scale compositional study of Iron Age pottery from Jerusalem was carried out by the author recently and may shed new light on the interpretation of the figurine production. About 70% of the figurines were produced with the Rendzina soil type (Figure 6; Ben-Shlomo 2019, Figures 7.36–7.38); yet, this was not the most common source used for the pottery vessels, with only 14% of the assemblage. It should be noted that there was a preference for the terra rossa soil type for the production of cooking pots and of Moẓa clay for that of storage jars, possibly creating a bias. For example, if only bowls are considered, 27% of these were produced from rendzina soil type clay (Ben-Shlomo 2019, Figure 7.40). Yet, still, in the background of these results, there seems to be a strong preference for rendzina calcareous soil for the production of figurines, which may indicate a specialized workshop or workshops producing that selection. The variability in temper and coarse inclusions can also be caused by the natural variability in clay sources.

On the other hand, as the other clay types (terra rossa and Moẓa marl) were preferred by pottery workshops in Jerusalem, the selection of the less popular clay type may indicate a household production, and this clay is highly available near the residential locations of the town, especially in the City of David (to the south and east, in the Qidron Valley) and the Ophel areas (to the east, see Figure 8: Menuha F.). Thus, non-professional potters (or different potters than those making regular pottery) produced the figurines, resulting in a different and less homogeneous clay composition. Therefore, whether the figurine production can be defined as “household” production is still an open question.

It is also possible that “lower-quality” clay was used for the figurines, or less energy was invested in clay collection and preparation and homogenizing the clay by the regular workshops potters, as figurines were not marketable as pottery vessels, and the properties of the clay was therefore less valued (for that reason they also may have been fired at a lower temperature). (Darby 2011, pp. 217–20) also suggested that Cave I in Jerusalem near the City of David (Holland 1995) may have been a location of a potter’s workshop producing figurines: the cave contained bone tools, extensive wet clay deposits, piles of broken potsherds, installations, and caches of pottery and ceramic objects including hundreds of anthropomorphic and zoomorphic figurines). Future petrographic analysis of this assemblage could put this proposition to trial, examining whether the figurative objects are made of more homogenous clay recipes and whether they match the clay deposits and pottery found in that location.

Another option is that the compositional characteristics of the figurines (and other cultic items) may be related to certain rituals or symbolic aspects associated with figurine production, and thus, the clay used might have had certain symbolic or other assumed advantages. Clay and the potter’s work are often associated as a metaphor of creation in the Old Testament (e.g., Jer. 18:6, see Darby 2011, pp. 411–33), and in Ugaritic and Mesopotamian texts (see Darby 2011, p. 58; clay is seen as a material that allows magic and transfer between media). This meaning could affect the actual process of making human- or animal-shaped figurines (or other cultic clay objects), and the selection and handling of the clay can contribute significantly towards the symbolic meaning of the object produced. Mesopotamian texts mention the making of figurines of various materials, including clay, in relation to certain, mostly apotropaic, healing rituals (Darby 2014, pp. 99–106, more references therein). These include a large group of Neo-Assyrian texts dealing with figurine rituals (Darby 2011, pp. 130–64). Whether these descriptions actually reflect specific treatment of the clay (mixing, levigation, or tempering for example), which, in turn, may be reflected by its analyzed composition, is not clear.

For example, various rituals involve creating clay figurines and using them in ancient Mesopotamian anti-witchcraft texts for example (Maqlû) (Abusch 2002; Darby 2011, pp. 143–45). In the manufacture of the figurines, the purification of the “clay pit” is often mentioned (such as by throwing flour). Other ingredients, such as wax, fat, ox blood, flour, donkey urine, reed, straw, excrement, or wood could be mixed with clay (e.g., Abusch 2002, pp. 34, 49; Darby 2011, pp. 147–49); yet, the figurines are usually disposed of in a jar or pit (as in “šep lemutti ina bīt amēli parāsu” texts [“to block the entry of the enemy in someone’s house”], see Darby 2011, pp. 150–57, more references therein). The figurines were probably made from molds prepared beforehand. The ritual includes instructions describing how to place them, paint them, etc. For clay figurines, purification of the clay pit is mentioned, and the clay was thence taken to the city, possibly to a potter who would make the figurines. Eventually, they were buried in a domestic household. These figurines may not, however, be visible in the archaeological record as they are not fired and are disposed of outside the household. Based on this background, it is possible that some of the figurines recovered in Jerusalem (and elsewhere) were manufactured at home or by priests involving similar rituals. Darby (2011, p. 86) suggested that the figurines depicted “lower level” divine or supernatural entities with apotropaic functions.

The compositional analysis of the cultic objects from Yavneh and Jerusalen may be an opportunity to compare cults and religions in Philistia and Judah during the Iron Age II. It may not be a coincidence that the Philistia example represents a temple assemblage, while the Judean one a domestic cult assemblage. Temples and shrines in Iron Age Judah are few, and at the few excavated, such as at ‘Arad, Moẓa, and possibly Khirbet Qeiyafa, cultic clay items are rare and have not yet been analyzed. On the other hand, no large-scale compositional analysis of Philistine domestic cult objects such as anthropomorphic and zoomorphic figurines or zoomorphic libation vessels has been carried yet.

5. Conclusions

The compositional analysis of Iron Age cultic items included two relatively large assemblages analyzed: a group of clay shrine models and other cultic items related to a temple at Yavneh, Philistia, and female pillar and zoomorphic figurines from domestic contexts at Jerusalem and Tell en-Naṣbeh in Judah. Both Yavneh and Jerusalem assemblages were found to be largely locally made at the sites they were found, while the figurines from Tell en-Naṣbeh were imported from the Jerusalem region. However, the temple-related assemblage showed higher standardization and may reflect a temple-related workshop. The domestic figurine assemblage may reflect household production or various local workshops. The assessment of the production mode was more viable, as large quantities or regular pottery vessels from the sites at Jerusalem and Naṣbeh were also analyzed. Clearly, many questions remain, yet, these studies, as well as others, show the potential of compositional analysis, especially petrography, for shedding more light on cultic and religious practices in ancient Israel and in general.