Mea culpa: revisiting the zooarchaeological perspective on the origins of metal use for quotidian activities at Jericho (Tel es-Sultan)

Introduction - foreword

This study is the story of science and how we continually push the envelope. As scholars, we really should not mind when we discover that we were wrong. In fact, this is what the scientific method is all about – testing and retesting and refining knowledge – hence the title of “mea culpa” – my fault.

Many years ago, I published what was then considered to be cutting-edge research on distinguishing metal from stone tool butchering marks at the site of Jericho/Tel es-Sultan. The goal was to find a proxy measure to identify when chipped stone tools are replaced by metal tools for such quotidian activities as butchering animals. Now, I know that my earlier study was flawed.

Over the past 25 years, I have endeavoured to examine the issue of the introduction of quotidian (daily life) bronze metallurgy to the Near East, Egypt, and Europe to gauge its effect upon social and economic developments. The spread of quotidian bronze metallurgy enabled broad segments of society to be part of the technological transformation from stone to metal.

Here, I revisit and reanalyse with improved methods the animal bone assemblage with butchering marks excavated by Kathleen Kenyon at the ancient city of Jericho/Tel es-Sultan and consider it in light of regional developments in the southern Levant. The resultant information is contextualised within the larger issue of the transition from a chipped stone to a hard metal based butchering technology across the region.

Alternative means to monitor adoption and spread of metallurgy

One alternative means to investigate the shift from a stone- to a metal-based technology has been through the study of lithic tool types. In the Near East, this has only been systematically attempted in the southern Levant where it was demonstrated that functional chipped stone tool types gradually disappeared between the ends of the Chalcolithic and Iron Age (Rosen, 1997). Some types disappeared because of changes in subsistence (arrowheads), while others are replaced with metal types that had a corresponding function (axes). Some stone tool types disappeared quickly (at or prior to the beginning of the Bronze Age - arrowheads), others disappeared more gradually (through the Bronze Age - axes), while some continue to be used into later periods (Iron Age - sickles). Rosen (1997) suggested that traditional materials continue to be used until a clear improvement in efficiency emerges. From his study, it is clear that the initial phases of the adoption of bronze metallurgy did not involve the wholesale replacement of flint tools. In other words, one cannot assume that new technologies will quickly replace older tried and true technologies, as is commonly assumed.

Another alternative to investigating the transition from display/prestige to quotidian technologies has been through the study of butchering marks on animal bones. After more than 30 years of research on butchering marks across the Near East (and southern Levant in particular), it has been demonstrated that there is little evidence for metal knife marks in the 3rd millennium BCE (Early Bronze Age/EB), and they are extremely few (De Sousa and Greenfield, 2023; Greenfield, 2005, 2013; Greenfield, 2021; Greenfield et al., 2018; Greenfield et al., 2021; Greenfield et al., 2022a, 2022b; Greenfield and Marciniak, 2021; Saidel et al., 2006). Metal knife mark identifications from the 3rd millennium BCE can almost be classified as background noise. The near dearth of metal knife marks from EB deposits corresponds with the fact that hard (10%) tin-bronze objects only begin to appear toward the end of the EB (Philip, 2001). It is clear from the numerous other analysed assemblages that metal slice marks on animal bones do not dominate across the region until the Late Bronze Age. By this time, the large-scale trade networks that allow copper and tin to be imported are well in place and such commodities are available, less expensive, and more accessible to the bulk of the population (Gailhard, 2008; Muhly, 1973; Rosińska-Balik et al., 2015; Williams et al., 2025).

Context

Copper metallurgy in the Near East (and southern Levant in particular) appears very early with the recovery of cold-hammered copper artefacts from periods as early as the Late Pre-Pottery Neolithic (c. 6000 BCE) (Mellaart, 1976; Redman, 1978). With the development of smelting techniques, copper metallurgy spreads relatively quickly throughout the region during the Chalcolithic (ca. 4700-3600 BCE (Levy and Shalev, 1989; Lovel and Rowan 2011; Rowan and Golden, 2009). Yet most of the studies of Chalcolithic copper items appear to be prestige or display items (Rowan, 1998; Stein, 1996).

The Bronze Age (BA - 3600-1200 BCE in the southern Levant) is distinguished by the alloying of copper with other metals (e.g., tin, arsenic or lead) (Adriaens et al., 2002). During this period, the new and harder form of metallurgy appears and spreads widely, especially for prestige items (Montanari, 2015, 2018; Nigro et al., 2018; Philip 1989; Rosenberg and Chasan, 2018; Walzer and Paz, 2023). In the southern Levant (modern Israel, Jordan, and the various disputed territories), hard bronze alloyed tools (10% tin) for quotidian activities begin to appear in the Early Bronze Age (EB II/3100-2900 BCE EB III/2900-2500 BCE in Israel), such as axes, adzes, chisels, etc., (Blackwell, 2011; Genz, 2000), but the vast majority are restricted to the latter half of the 3rd millennium (Intermediate Bronze/IB/2500-2000 BCE). By the Middle Bronze Age (ca. 1950 BCE in Israel), the use of bronze becomes more widespread and common, even for quotidian activities (Moorey, 1994). Yet the question remains as to when do the majority of people in such early complex societies begin to have access to and use bronze tools (i.e., knives) for butchering animals.

Daggers are morphologically very different – they have two slicing edges and tend to be symmetrical from one edge to the other, but are mainly used for stabbing. They are a fighting weapon. In contrast, knives have a single thin slicing edge while the other edge is much thicker to provide stability to the slicing edge. They are used for slicing (e.g., butchering).

Daggers appear earlier than knives in the archaeological record of the Levant and the Near East in general. They occur already toward the beginning of the 3rd millennium BCE, while knives appear much later in the same millennium (Arcan, 2012; Montanari, 2015; Nicholson and Shaw, 2000; Nigro et al., 2018; Philip, 1989; Zimmermann, 2004, 2005). Some of the earliest bronze knives in the southern Levant come from the IB and MB tombs of Jericho (e.g., Tombs M11, D22, and P19) (Kenyon, 1965, 237, figure 111; Montanari, 2014; Nigro et al., 2018) which is where much of the zooarchaeological data presented here derives.

Problems with relying exclusively on metal artefacts for understanding the adoption of metallurgy

The most common method to monitor the shift from a stone to a metal-based technology has been through the presence/absence and frequencies of the various metal tools preserved in and recovered from the archaeological record. Most studies concerned with the origins and spread of metallurgy are based on finds of metal artefacts recovered from archaeological contexts (Chernykh, 1992; Levy and Shalev, 1989; Moorey, 1988; Moorey, 1994; Muhly, 2011; Shephard, 1980; Tylecote, 1992, 1987, 1986). However, this approach has significant limitations that hampers a systematic evaluation of the nature of the spread of metallurgical technology except from the perspective of those items that have survived.

Only a relatively small percentage of what was probably a much larger metal assemblage survives to be found. Most such finds derive from a relatively restricted range of depositional contexts (e.g., hoards, burials) and relatively few come from settlement sites where they would have been used in daily life (Taylor, 1993).

In general, metal is one of the few raw materials that can be recycled. While copper-based metals do not rust, they still can deteriorate depending on soil conditions. More likely in antiquity, broken, discarded or no longer needed copper items could be recycled – melted and made anew. The result would be little surviving evidence to reconstruct ancient patterns of metal adoption and use.

In contrast, the number of chipped stone tools recovered from sites of the same period dramatically outnumber the frequency of metal objects (Greenfield and Marciniak, 2021). Far fewer metal tools have survived from ancient times in comparison to to chipped stone tools. This is not to imply that metal was less important, but it is likely an issue of both preservation and recyling.

Advantages of a zooarchaeological approach

There are several advantages of a butchering mark approach to the study of the spread of hard tin-bronze metallurgy in comparison to studying the actual implements. Both are important, but the number of metal implements is heavily reduced by recycling and decomposition of metal objects.

Butchering marks can be used to identify the type of implement used (e.g., knife, axe, saw, hammer, etc.) as well as the raw material of the implement (chipped stone, metal). Even the type of chipped stone raw material can be inferred (Aramendi et al., 2023; Greenfield, 2006; Greenfield and Marciniak, 2019; Maté-González, et al., 2018b; Yravedra et al., 2017).

Animal bones are ubiquitous on Near Eastern sites and when collected can range from hundreds to tens of thousands in many excavations (Hesse and Wapnish, 1985). By distinguishing whether butchering marks on animal bones are made by metal or stone tools, an independent measure of the relative importance of the different raw materials used for butchering cutting can be generated. The spread of metallurgy, as a result, can then be monitored both chronologically and spatially within and between regions, and within settlements from building to building (Greenfield, 2013, 2017; Greenfield and Marciniak, 2019; von Lettow-Vorbeck, 1998; Yravedra et al., 2009).

Thus, the two approaches (bone butchering marks vs. metal artefacts) yield different but complementary understanding of the shift from a lithic to a metal-based technology (see Eriksen, 2010; Greenfield, 2013; Hamon et al., 2024; Jones, 2011; Rosen, 1997; Vandkilde, 1996). They inform as to the types of technologies used, and who in a society is getting access to the new metal food processing technology. It shifts the focus from the elite to the larger population of society. The data can then become a proxy measure for the frequency (and thereby importance) of the different raw materials in the quotidian domain of food processing (animal butchering).

Jericho assemblage

Location and curation of Jericho assemblage

The site is located at the edge of the modern city of Jericho in the Palestinian Authority Area (Figure 1). The basic descriptive analysis of the entire zooarchaeological assemblage from Jericho was undertaken and reported on in several seminal publications elsewhere (Clutton-Brock, 1979; Clutton-Brock and Uerpmann, 1974).

All of the faunal remains analysed here are stored in the Natural History Museum, London. They were examined by the author during initial visits in 2000 and 2002, and reanalysed in 2022. Additional remains not included in this study were disbursed by Kenyon and colleagues to several academic institutions around the world (e.g., University of Sydney), as were the human remains and various artefacts. The original bone assemblage continues to be housed in the curatorial facility of the Natural History Museum of London (BMNH).

Problems with previous analysis of the assemblage

There are several problems with my previous analysis of the Jericho specimens: By lumping the specimens into two broad categories (EB and MB), the finer chronological distinctions evident in the data were lost.

• It was conducted almost 25 + years ago when I was just developing the method;
• I lumped together into a single category all the specimens from EB contexts (EB I, II, III, IV/IB) without consideration of possible changes over time within the EB;
• Similarly, I lumped all the MB specimens into a single category; and
• I did not have the level of microscopic technology available to me during the analysis of the moulds of butchering marks.

It also led to mistaken conclusions. A few metal slice marks were mistakenly identified on some the pre-metal age specimens in the earlier report –BMNH #6241 and 6237 (Greenfield, 2005). This is corrected here. These specimens have been reanalysed and correctly listed as chipped stone specimens in this analysis.

Another error is that the results of the earlier data analysis mistakenly suggested that metal knife butchering marks were evident throughout the EB (Greenfield, 2005). It inflated the importance of the EB over developments in the subsequent MB. This led to the misconception that bronze metallurgy for butchering appeared earlier in Jericho than elsewhere in the region and in much greater frequencies (Greenfield, 2013, 2021).

Another misconception stemming from the lumping of MB remains is that it ignored the changes occurring during the MB. The very high frequency of metal butchering marks in the MB at Jericho always stood out as unique when compared with the frequency of metal vs. stone slice marks from other contemporary sites from the Jordan Valley. However, more recent studies of butchering marks on sites with finer time control in the region (e.g., Tal Zira'a) have shown that metal marks only begin toward the end of the IB, and are increasingly represented among butchering marks through time during the MB II (Greenfield et al., 2022b). Given the methodological and temporal issues with the earlier Jericho study, it was deemed that the Jericho assemblage required a complete temporal and methodological reanalysis to account for the discrepancy between the pattern at Jericho and contemporaneous sites that displayed a different pattern of metal adoption for animal butchering.

Chronology of the assemblage

Several major and minor periods of occupation were defined at Jericho by Kenyon in her excavations of the Early and Middle Bronze Age. The earliest are the EB I-III deposits from the Tel. However, the chronological relationships of the published stratigraphy from the Tel are very problematic.

All the bones in the Natural History Museum (BMNH) are labelled with the museum's catalogue numbers. Juliet Clutton-Brock made a detailed list of each of the bones and their associated depositional context known at the time. There is a large-scale printout in the museum's archives with the detailed provenance of each bone. Further, the bones are laid out in a series of drawers where they are organised by time period.

Kenyon published detailed section drawings of the baulks, but there is no concordance (at least that I was able to find) that links the excavated animal bones from the Tel with Early Bronze (EB) I-III or Middle Bronze (MB) IIA-C stratigraphic phasing. The EB I-III material from the Tel are simply labelled as EB. The same is true of the MB IIA-C material from the Tel – simply labelled MB in the Natural History Museum. Thus, it is impossible to obtain a finer chronology of the EB and MB specimens from the Tel.

The situation is very different with respect to the tomb groups that were excavated off the Tel. Kenyon discovered and excavated tombs in several off-Tel areas. She published the tomb groups in great detail along with their chronological relationships (Kenyon, 1960, 1965). These can be divided into several chronological periods and subperiods, including Intermediate Bronze (IB) (general, early, and late), MB (general, IIA, IIB, and IIC) and are used as the basis for identifying change over time at the site in the use of butchering technology.

As the goal of the study was to identify the changes in technology used for butchering with the introduction of bronze, butchered remains from later time periods (e.g., Iron Age and later) were not systematically analysed. One note of relevance is that the faunal material from the Byzantine period was examined, but no butchering slice marks were found, only metal saw and chop marks. These are not described here.

The data from the pre-Bronze Age periods are included in this study as the goal was to ensure that metal butchering marks did not begin in a time period long before the metal ages. Otherwise, this would have undermined the validity of the study.

Thus, the faunal remains discussed in this report recovered during the 1952-1958 large-scale excavation at the site of ancient Jericho/Tel es-Sultan conducted by Kathleen Kenyon can be divided into five phases for this analysis (Table 1). The absolute dates follow the high resolution chronology recently developed for the southern Levant (Greenberg, 2019; Lev et al., 2020, 2022; Regev, 2013; Regev et al., 2012a, 2012b, 2020, 2021). These dates are not yet accepted by all (Montanari, 2014, e.g., Nigro et al., 2018, 2019).

Table 1. Early and middle bronze Age chronology in the southern levant (BCE).

Analytical methods

To identify bones with butchering marks on them, each bone fragment in the Natural History Museum Jericho Collections was individually examined in the curatorial facility of the museum. Most of the butchering marks were readily identifiable to the naked eye. A low power magnifying glass was used to examine to enhance the potential for discovering butchering marks. A binocular microscope was further used to further identify microscopic grooves on bones (10–40x).

The presence of butchering marks has long been recorded in the zooarchaeological literature (e.g., Ashbrook, 1955; Guilday et al., 1962; White, 1952b, 1952a, 1954, 1953, 1955). It has long been recognised that the nature of butchering activity leaves a relatively clear and different morphological signature than tooth and other taphonomic causes of grooves (Binford, 1981; Fernández-Jalvo and Andrews, 2016; Gifford-Gonzalez, 2018; Shipman, 1981). Butchering marks can even be identified by the naked eye – as they are typically relatively short straight incisions in the bone (See Figure 2).

Multiple butchering slice marks at the same location create intersecting grooves. In these cases, the later strokes fall into the original groove and modify it. They destroy the morphology of the previous slices, rendering it often difficult to determine whether they were made by stone or metal instruments (Figures 3a, b). What may look like a metal knife mark by eye (Figure 3) or in a light optical microscope may in fact be a chipped stone tool when viewed in a SEM (Figure 4).

Many bones with potential slice marks were rejected when subjected to microscopic examination since they were deemed to be caused by non-butchering sources. Some grooves that were initially thought to be butchering marks were ultimately rejected as being made by other taphonomic sources (e.g., conservator marks, gnaw marks, trampling, etc.) (cf. Andrews, 1995; Backwell et al., 2012; Behrensmeyer et al., 1986; Gifford-Gonzalez, 2018; Lyman, 1994; Olsen and Shipman, 1988; Shipman, 1981).

Once butchering marks were identified, the bone and the butchering marks were drawn to scale (1:1) and subjected to a detailed zoological analysis (including taxon, state of domestication, symmetry, part of the element, age-at-death, etc.). Almost all of the bones (and fragments) with butchering marks were identified to a specific taxon, at least to genus (and usually to species), and to element. Very few taxonomically unidentifiable specimens are in the assemblage as the excavators appear to have selectively recovered the more identifiable specimens. It is likely that only taxonomically identifiable bones were preferentially collected and kept by the excavators. The taxonomic identifications made in this study for the most part agreed with those of the earlier analysts (Clutton-Brock, 1971, 1979; Clutton-Brock and Uerpmann, 1974), but occasionally differed. These were noted in the museum's registry.

Next, slice marks were examined under a light optical microscope and a tentative identification of the nature of the raw material of the implement was made – whether it was made by a chipped stone or metal slicing tool.

Subsequently, silicone moulds (made from a light body polysiloxane silicone material) were made of the groove in the bone to enable further study the mark in a SEM. After curing, each mould was examined to ensure that an accurate copy of the groove was obtained (Rose, 1983). One has to be careful at this point to ensure that no air bubbles were trapped in the silicone particularly along the axes of the slice mark (Figure 5). Occasionally the groove was too narrow for the silicone to penetrate which created air pockets along the apex (visible as holes or concavities in Figure 6). These will prevent analysis of the moulds itself in a later stage.

In general, the assemblage was clean of any calcium carbonate concretions. This made it much easier to identify marks than in many other contemporary assemblages from the region.

Assemblage conservation

One problem with the Jericho assemblage is that most of the assemblage was conserved at one point at the museum (based on a November 1972 note in the archives) in a solution of 50% PVA and 50% of Dis H₂O. Sometimes, the PVA was so thick that it appeared as if the bone had a wax-like covering (87% of specimens show some kind of preservative on them – Table 2) In many cases, the PVA/waxy substance was so thick that it filled the groove and a mould could not be made of the butchering marks. Despite the above issues, moulds could be taken of the vast majority of specimens as the PVA did not enter the groove and analysis could take place (71% - Table 3).

Table 2. Frequency distribution of butchered bones treated with preservative by period and subperiod (bNISP).

Bold value indicates a sum or subtotal for the data categories.

Table 3. Frequency distribution of silicone moulds of butchered bone slice marks by period and subperiod (bNISP).

Bold value indicates a sum or subtotal for the data categories.

In several of the SEM photographs, it is evident that the silicone mould did not penetrate to the bottom of the groove to capture the entire slice mark correctly. Most of the time, it was because PVA filled the apex of the groove.

In cases where moulding was not possible, the preservative was often sufficiently transparent to be able to see the morphology of the butchering mark and make a visual identification of the technology used in the butchering process with a light optical binocular microscope.

Silicone moulds were not made of most chop marks as the chop marks usually only preserved one side, and that was often damaged. The exception is BMNH cat. No. 6191 (SEM Specimen 284) where what seems to be a chipped stone axe marks were identified on a Bos taurus proximal humerus.

Four Pottery Neolithic bone fragments that come from Jericho context O200/19 that were listed in the museum catalogue with butchering marks were not found in the assemblage by the analyst. In this case, the specimens are included in the analysis even though moulds were not taken.

Transportation and curation of moulds

The advantage of silicone moulds is that they are easily transportable, are flexible and not easily damaged, make microscopically identical copies of the original surface being moulded, and are not subject to the limits of local and international heritage/antiquity laws. Thus, a three-dimensional accurate copy of the original groove can be examined in laboratories that house the necessary technology. However, one must be careful particularly with moulding of bones where the hard cortical surface has been damaged and where the cancellous structure has been exposed as the moulding material may penetrate into the bone structure and cannot be easily removed (Rose, 1983; Valtierra et al., 2023).

A further advantage of silicone moulds is that they produce a reverse image of the slice mark. Instead of having to peer deep into a groove in the bone, the groove is reversed. It stands up like a mountain. It is possible to observe important diagnostic criteria (such as apex symmetry and lateral striations) that are essential for making an accurate diagnosis. This is evident in the SEM photographs.

The moulds were exported to the Near Eastern Biblical and Archaeology Lab of St. Paul's College (University of Manitoba, Canada) where they were examined with a binocular light optical microscope to confirm the initial identification of the raw material (stone or metal) used to make the slicing mark and to determine if it was suitable for further analysis in the lab's SEM (Cambridge Instruments SEM Stereoscan 90 and Zeiss EVO Series 10 SEM). This ensured a level of accurate repeatability of results rarely encountered in such analyses. The moulds are curated at the University of Manitoba in the Near Eastern and Biblical Archaeology of St. Paul's College and available for reanalysis.

There are many advantages to using a SEM. First and foremost is the high resolution, followed by the depth of field (Rose, 1983; Shipman, 1981). Most light optical microscopes (including Dino-Lites with ring lights) are designed to examine flat specimens on a horizontal plane and thereby flatten the image. In contrast, the SEM provides an image that is very accurate with a high depth of field. As a result, morphological diagnostics can be identified and quantified to distinguish between chipped stone and metal slicing tools used to butcher animals.

Slice mark identification criteria

There has been more than 40 years of research on the different sources for grooves on bones. Most of the research has focused on Palaeolithic assemblages to distinguish between the cultural and natural forces that create grooves on bone. As a result, it is possible to distinguish grooves made by butchering marks from other attritional forces, such as carnivore tooth, trampling, weathering and the like (Arriaza et al., 2025; Cifuentes-Alcobendas and Domínguez-Rodrigo, 2019, 2021; Courtenay et al., 2019; Domínguez-Rodrigo, 1999; Maté-González et al., 2016; Maté-González et al., 2018b; Moclán et al., 2018; Olsen and Shipman, 1988; Potts and Shipman, 1981; Shipman, 1981; Val et al., 2017).

In recent years, new methodological approaches to the identification of slice marks have been successfully applied – in particular geometric morphometric and machine learning modelling (Aramendi et al., 2023; Courtenay et al., 2018; Maté-González et al., 2018a, 2018b; Maté-González et al., 2023). In general, they support the traditional identification criteria for distinguishing metal from stone chipped stone tool slices developed over the past four decades.

In this study, the concern was whether it is possible to distinguish between metal and chipped stone slicing marks. The analytical method for distinguishing between chipped stone and metal tool slice marks is based on experimental research conducted by the author and others (above). The diagnostic results are discussed at length elsewhere (Domínguez-Rodrigo et al., 2017; Greenfield, 1999, 2002, 2006; Maté-González et al., 2017; Olsen, 1988a, 1988b) and are only briefly summarised next. While experimental research has also been conducted on identifying metal and stone tool axe marks (Okaluk and Greenfield, 2022), the focus here is on slice marks as very few chop marks were identified in the assemblage.

Metal knife blades produce a very uniform pattern, which can be summarised as follows (Figures 7E, F):

• Produce sharp V- or hard cornered |_|-shaped (when dull) groove that meet at an apex at the bottom of the groove. The groove stands up in the silicone mould;
• The apex of the groove tends to be symmetrical regardless of the angle that the tool entered the bone;
• Since metal tools need to be sharpened on both sides, they produce a uniform set of patterns on both sides of the groove;
• Considering the resharpening process, there are no lateral striations (seen often as steps in the SEM); and
• In general, metal knives produce a cleaner and more even groove. The only difference is when modern saw-like knives with a serrated-edge blades are introduced and they produce more of a saw-like groove.

Chipped stone slicing tools can be divided between nature of production (unifacial vs. bifacial) and retouch (unretouched, unifacial, and bifacial). It is not possible to distinguish between unretouched blades and flakes. The morphology of the groove is identical. In contrast, retouching produces diagnostics that are identifiable (Figures 7A–D).

• In general, chipped stone tools produce a shallower, less even and more uneven slice mark than metal slicing tools;
• Unifacially produced unretouched chipped stone tools produce an asymmetrical (uneven, almost lopsided) groove that is not reflective of the tool's angle of entry into the bone;
• Chipped stone tools produce an apex in the groove that will be to one side of the groove;
• Both unretouched chipped stone tools blades and flakes (who grooves are otherwise indistinguishable) have a ventral and lateral face that create distinctive grooves. The ventral face will create a steeply rising and smooth surface in the groove. The lateral face creates a face that rises more gently
• On unretouched tools, the uneven lateral face of the tool will usually have at most 1–2 steps or lateral striations on the gradually rising face. These generally run parallel to the apex of the slice mark, depending on whether it is retouched or not;
• Unifacially retouched tools will have many lateral striations on one side, while bifacially retouched tools will have many lateral striations on one or both sides of the apex; and
• Scrapers can also be used as slicing tools, but they appear to produce very uneven, broad and wavy marks (Figure 7D).

Quantification

The data were recorded with two different quantitative measures: NISP and bNISP. The former (Number of identified specimens) is used for the assemblage in general. It is normally assumed that each bone represents a different individual. However, if two or more bones are articulated (thereby representing the same individual), then the group of bones would receive a NISP of 1. While NISP is useful when quantifying by taxa in order not to inflate species counts, it is very susceptible to the influence of differential fragmentation and may overrepresent what was there. But it has been found most useful in quantification of taxonomic representation in a number of quantitative studies (Grayson, 1984; Lyman, 1994, 2018; Marshall and Pilgram, 1993).

The latter measure (butchered NISP) is used for only the butchered data. bNISP counts the number of specimens with butchering marks. This is a subset of the larger NISP (Beller et al., 2022; Greenfield et al., 2016; Greenfield et al., 2018; Greenfield and Brown, 2016).

Taphonomy

A number of processes must be considered when examining and quantifying an assemblage as the analytical assemblage may be quite different than the original assemblage due to weathering, burning, and other attritional processes (Efremov, 1940; Fisher, 1995; Lyman, 1994).

The bone surface is often affected by the degree of weathering. Weathering occurs when the bones are left exposed on the ground surface for prolonged periods of time before burial or when re-exposed after burial. Smaller mammal bones weather more quickly than those of large mammals. Weathering patterns are divided into four categories, fresh (no cracks or flakes), light (small cracks, mainly parallel to fibre structure), medium (flaking and loss of bone surface, cracks enlarging) and heavy weathering (considerable bone surface loss, splintering, falling apart structurally). The definitions follow those outlined elsewhere (Behrensmeyer, 1978; Greenfield, 1986).

Specimens were coded for whether they displayed evidence for burning. It is likely that most bones from the Tel and tomb deposits were there because they were cooked. If they were cooked in vessels or were still encased within the meat when over a fire or in an oven, they would not display evidence of burning (charred or calcined). In contrast, bones that were cooked in stews or were encased in meat and not directly exposed to a flame or heat source could be treated as boiled. Unfortunately, evidence for boiling and baking were not systematically recorded during the analysis (Buikstra and Swegle, 1989; Greenfield and Beattie, 2017; Koon et al., 2010; Shipman et al., 1984).

Canid and rodent gnawing are another source of bone attrition commonly found in many sites in the region. The type of mark made by each taxon is very different and relatively easy to identify (Atici, 2006; Brain, 1969; Gifford-Gonzalez, 2018; Greenfield, 1988).

Results

Butchered assemblage size by time period

The faunal remains recovered from the site derive from many deposits that extend from the Protoneolithic to the Middle Bronze Age (Table 4). The remains in the assemblage that did not display any butchering marks were not analysed nor described here. This report details only those periods from which butchered (i.e., those with slicing marks) remains were identified.

Table 4. Frequency distribution of butchered and all bones examined in the assemblage by period and subperiod (bNISP of butchered and NISP of all bones).

Bold value indicates a sum or subtotal for the data categories.

This phase of the study of the Jericho faunal remains also discovered more remains with butchering marks than were initially identified in the earlier report on the assemblage (Greenfield, 2005). These are now included in the analysis. In the previous report, a total of 271 bones were found with evidence of butchered remains (Greenfield, 2005). In the updated analysis described here, 382 specimens were identified with butchering marks out of a total of 583 specimens examined in the assemblage (Table 4). This is an increase of almost 30% of specimens which allows some patterns to appear that were not evident previously.

The data are divided between major periods and subperiods, and the frequencies are very uneven. Major periods include specimens from the Protoneolithic (n = 4), Pre-Pottery Neolithic A (PPNA: n = 8) and B (PPNB: n = 19), Pottery Neolithic (PN: n = 4), EB (n = 17), IB (n = 69), and MB (n = 248, 65%) (Table 4). Clearly the frequency of MB specimens dominates the assemblage. In many respects, this is a positive development as the MB has now become the major period when the adoption of metallurgy for butchering begins to become significant (Greenfield et al., 2022b).

Very few specimens come from the various other periods. The next largest, in descending order come from IB (18%), EB (4.5%), PN (1%), PPNA and PPNB (4.7% and 5%, respectively), and Protoneolithic (1% - Table 4).

Slightly more than half of the IB and MB butchered bones derive from tombs (51%). The rest came from Tel deposits (49% - Table 5). These frequencies mirror the overall distribution of tomb vs. Tel deposits in the MB. All the other specimens derive from Tel deposits.

Table 5. Frequency distribution of butchered bones by period, subperiod and depositional context – Tel or Tomb (bNISP).

Bold value indicates a sum or subtotal for the data categories.

While the EB material could not be subdivided into subperiods (i.e., I, II, or III), the IB (early, late) and MB (IIA, IIB, and IIC) could often be divided between each of their subperiods. Those specimens that could not be assigned to a specific period are listed as “general” in the various tables.

Butchered assemblage taphonomy

It is essential to understand the taphonomy of any assemblage as various forces (natural and cultural) can create marks on bones that can be confused with butchering marks. Some of these are discussed next.

Weathering

The vast majority of the assemblage is very lightly weathered (80%), as the material comes from either deeply buried Tel deposits (Protoneolithic-EB), or tombs (IB and MB – Table 6). The frequency of lightly weathered material is even greater when just the MB material is considered (86%). This is not the case when the IB material is considered as less than half is lightly weathered (43%) and more than half are medium weathered (55%) and one fragment was heavily weathered (1.4%). The IB weathered distribution is very unusual when compared to the MB as it is mostly tombs and one would expect a higher degree of preservation.

Table 6. Frequency distribution of butchered bone by period, subperiod and degree of weathering (bNISP).

Bold value indicates a sum or subtotal for the data categories.

Gnawing

In general, most of the bones in the assemblage do not display any evidence of gnawing (Table 7). Only four specimens (1.05%) displayed any evidence of gnawing – two were from IB and 2 were from MB contexts.

Table 7. Frequency distribution of gnawed and not gnawed butchered bone by period and subperiod (bNISP).

Bold value indicates a sum or subtotal for the data categories.

Several gnawed specimens (n = 4) were originally thought to have been slice marks made by stone (n = 1) and metal (n = 3) slicing implements. Reanalysis of these demonstrated that they were simply gnawed specimens. These have been corrected in this study.

Evidence for heating

Only a small percentage of the overall assemblage had burn marks on them, suggesting that they had been roasted over a fire (Table 8). They came from a variety of periods, and the frequency was highly variable. The highest frequency came from the PPNB (36.84%), followed by the MB IIA (17.5%) and PPNA (16.7%). In general, the relative frequency between periods does not follow any particular pattern.

Table 8. Frequency distribution of burned butchered bone by period and subperiod (bNISP).

Bold value indicates a sum or subtotal for the data categories.

The exception to this pattern is the MB where most of the burned bones derive from tombs rather than the tell (n = 12 vs. 1 specimen) (Table 9). Nonetheless, most bones from either tomb or tell did not display any evidence of burning. One can surmise that the burned bones in the tombs were offerings often roasted over an open flame, while those from the settlement area were heated/cooked in other less obvious means, such as boiling that may have been subsequently hidden by the preservative applied to the bones.

Table 9. Frequency distribution of burned and unburned butchered bone by period, subperiod and depositional context (bNISP).

Bold value indicates a sum or subtotal for the data categories.

Technological analysis

Chop vs. slice marks

Most butchering marks on the bones come from knife slice marks (95.8% - Table 10). Most specimens that displayed slicing marks derive from the MB (n = 236). Only four were from the Protoneolithic and PN, 17 each from the PPNA and PPNB, 16 from EB Tel, 69 from the IB tombs, and 236 from the MB tombs and Tel deposits (most of which come from the tombs).

Table 10. Frequency distribution of butchered bone by period, subperiod and butcher type (bnisp).

Bold value indicates a sum or subtotal for the data categories.

Only sixteen chop marks (4.2%) were identified. A few came from the PPNA and PPNB, and one from EB (SEM Specimen 5, BMNH #H450-36) deposits. Most of the chop marks (9 out of 12) came from the MBA tombs, while the rest came from MB Tel deposits.

The paucity of chop marks should not be interpreted to mean that more slicing than chopping was done at the site. In the first place, diagnostic criteria for chop marks were not clearly defined when the original study was conducted. Second, the problem with chopping activities, either from an axe/adze or other heavy implements, is that it usually fragments the bone. In effect, it destroys evidence of its own existence. All that is usually left are a variety of fragments, only some of which may preserve the impact mark. It is only with the advent of hard tin-bronze axes that we begin to see bones being clearly chopped. Before this moment in time in other sites in the region, most bones are broken by being bashed with a blunter instrument (Greenfield et al., 2022b). Unfortunately, while bash marks were evident on the assemblage, they were not systematically recorded as the diagnostic criteria had not been worked out at this point.

In contrast, slicing marks usually do not destroy or fragment the bone to the same extent. As a result, there is better preservation of the distribution of slicing marks on bones.

Stone vs. metal slice marks

During the 2022 microscopic reanalysis of the silicone moulds taken in earlier years, a number of specimens was deemed to be poor candidates for the determination of metal or stone tool marks (n = 147). The slice marks were filled with PVA and were ultimately indeterminate. Therefore, this section of the analysis is based on only 237 specimens (Table 11).

Table 11. Frequency distribution of butchered bone by period, subperiod and butchering technology (bNISP) based on reanalysis of silicone moulds.

Bold value indicates a sum or subtotal for the data categories.

The PPNA and PPNB chop marks were for the most part created by stone tools of various types (axes and hammers). The single EB chopped specimen was different and displayed a morphology expected with metal axes – very flat and straight sides.

Among the slice marks, both stone and metal marks were identified in the assemblage (Table 11). All the Protoneolithic, PPNA, PPNB, PN, and EB specimens were made by chipped stone tools (Figure 8).

During the IB, almost all of the specimens display the characteristics of chipped stone tool slicing marks (Table 11). Each specimen demonstrates the typical patterns described above for unifacially produced unretouched chipped stone tools. Only a single IB specimen was tentatively identified as having been made by a metal knife blade tool (SEM Specimen 138, Tomb O4; BMNH #5430). However, its temporal context is uncertain as the notes on the specimen in the museum accord it to an MB tomb, while Kenyon lists it as an IB tomb (Kenyon II, 116 ff.). Some tombs continue in use from the IB into the MB which may account for this early appearance.

Specimens with butchering marks that were made by chipped stone tool continue to exist in the MB (Figures 9, 10; Table 11). In general, they display the typical characteristics of grooves made by unifacially produced unretouched chipped stone tools. It is during the MB that dramatic shifts in the nature of butchering technology begin to appear. It is in the MB that the transition from an exclusive reliance upon chipped stone to metal slicing tools occurs when the data for the entire MB assemblage is pooled (69% chipped stone and chipped stone? total vs. 30.2% metal total for the MB in general - Table 11).

When the butchering data within each MB subperiod are examined separately, a very different story emerges. In the MB IIA, most specimens were butchered with chipped stone tools (70.27% chipped stone and chipped stone? vs. 29.73% for metal). In the MB IIB, the situation is reversed. For the first time, more specimens were butchered with metal tools than with chipped stone tools (27.78% chipped stone vs. 72.2% metal - Table 11). However, the situation reverses itself once again in the MB IIC where 90%–100% of the butchering marks were made by chipped stone tools. There is little to no evidence of metal slicing marks in this final subperiod of the MB.

While the small absolute frequencies from the various MB II subperiods make any firm conclusion difficult, the large frequencies for the combined MB subperiods demonstrate that metal tools began to be used, but were not the dominant technology. Further, the quality and hardness of the metal implements must have been relatively low as none of the metal butchering marks appear to be made by very sharp and hard instruments (Figures 11, 12). This back and forth pattern parallels the data from other sites, such as Afek and Zira'a. The MB appears to be the time period when the new technology is being adopted, but the older technology is by no means quickly or even fully displaced. The MB is probably best characterised as a period when both raw materials are used for tools used in the butchering process (Beller and Greenfield, 2026; Greenfield et al., 2022b).

When the data from MB Jericho are broken down by depositional context, there are temporal and contextual differences between the tomb and Tel materials with respect to whether metal or stone tools were employed in the butchering process (Table 12). From the specimens assigned specifically to the MB IIA, IIB, and IIC, all come from tombs. There is no contextual variation.

Table 12. Frequency distribution of butchering technology by period, subperiod and depositional context (bNISP).

Bold value indicates a sum or subtotal for the data categories.

In the MB IIA, metal slicing marks dominate (70.27% chipped stone vs. 29.73% metal). This changes in the the MB IIB when begin metal butchering marks come to dominate (72.2% vs. 27.78% chipped stone). The trend reverses itself once again in the MB IIC as chipped stone tools dominate the assemblage (100%). However, the MB IIC assemblage is very small (only four specimens).

Context becomes important when the unknown MB subperiod specimens are considered. Specimens from the Tel are dominated by chipped stone tool marks (67.7%+10.8%=78.5% chipped stone vs. 21.54% metal) (Table 12). A much smaller frequency of metal butchering marks derives from Tel deposits (21.5%). This may suggest that stone tools continued to be the dominant tool type for quotidian activities in living spaces on the Tel, as opposed to activities associated with the afterlife and the land of the dead in the tombs. The four specimens from the tombs are all uncertain chipped stone tool type designations.

Discussion

In an earlier version of this study, the chronology of the remains was less refined and the data were not clearly separated by depositional context (Greenfield, 2005). As a result, the developments seen in the IB at Jericho were subsumed within the general EB sample of remains. The data then appeared to suggest that metal knife butchering marks appeared in the EB. However, recent studies of contemporary assemblages from the region suggested that a different pattern existed. It was discovered upon a review of the comparative and Jericho data that the chronological divisions were too gross (EB vs. MB) to be able to identify the shift in technology. Finer temporal subdivisions were necessary.

In this more recent study, more specimens have since been located and their depositional context taken into consideration. It is now recognised that metal slicing marks in significant frequencies only appear very early in the MB, pushing forward the adoption of metal knives for quotidian purposes by almost a millennium. Also, context is now demonstrated to have an influence on any understanding of the use of stone vs. metal butchering tools.

In this more updated study, the discovery of the total and near absence of metal tools in the EB and IB is not surprising given the nature of copper-based metallurgy in these periods. Most copper-based objects across the Near East are low tin-copper or arsenical alloys (Hauptmann et al., 1993; Hauptmann and Pernicka, 2004; Odler, 2016; Tadmor et al., 1995). Only copper or occasionally natural arsenical alloys (i.e., arsenical bronze) are in use until the very end of EB III when tin makes an appearance (Shalev, 1994; Tylecote, 1992). The appearance of hard tin-bronze alloys (10%) appears in the IB, but only begins to become common in the MB. A similar process occurs in Mesopotamia (Moorey, 1994). The evidence from the Jericho butchering mark analysis conducted here supports such a contention. A functional butchering bronze metallurgy is entirely or almost non-existent until hard bronze alloys become widely available. Thus, the data from Jericho are important for increasing our understanding of the spread and rate of adoption of a functional metallurgical butchering technology across the southern Levant. It would appear to be adopted in spurts, similar to the process described for the abandonment of stone tools (Rosen, 1997).

Chipped stone tools (specifically flakes and blades) as the primary butchery tool are progressively replaced over time in the Middle and Late Bronze Ages by metal knives (Greenfield, 2004, 2013; Greenfield et al., 2018; Greenfield and Brown, 2016; Greenfield and Greenfield, 2018; Saidel et al., 2006). Similar studies conducted in Egypt have demonstrated that chipped stone slicing instruments continue to be utilised through the end of the Early Bronze Age and into the 2nd millennium BCE Middle Kingdom in Egypt (De Sousa and Greenfield, 2023; Graves-Brown, 2015, 2010; Luff, 1994). This is not surprising given that hard tin-bronze was likely a rare commodity, and the constituent parts were imported and transported over great distances across the Near East (Muhly, 1985, 1998, 1993; Özbal and Aslihan Yener, 1987; Vandiver et al., 1993). Therefore, hard tin-bronze would be an expensive resource, was limited in its distribution to elites, and likely used primarily for display and ritual (non-quotidian) activities, such as warfare, status enhancement, ceremonial, and ritual purposes.

Conclusions

A major problem that continues to affect this more updated analysis of the Jericho butchered assemblage is the low sample sizes for many of the time periods. It is difficult to obtain a statistically significant pattern as a result. This problem can only be rectified through the collection of larger samples from newer excavations at the site. Nonetheless, the data from Jericho allows us to draw some tentative conclusions.

While the data from Jericho are admittedly very small and were collected haphazardly in the field, the sample still represents a significant data-set for many of the periods in the region. Few other sites have each of these periods and depositional contexts represented in their assemblages.

The evidence from Jericho reevaluated here indicates that there were changes in butchering technology over time, but at a later period than originally reported. The change is not in the IB (or EB in general). Instead, it is during the MB, with the frequencies going back and forth between stone and metal during the various subperiod. The EB is really an extension of the Stone Age patterns seen previously. The pre-Bronze Age and EB assemblages rely exclusively upon stone tools for such quotidian activities as animal carcass butchering. The tools were for the most part unmodified flakes, haphazardly made, or simple unretouched blades. In contrast, the MB data indicates a substantial shift in butchering technology from stone to metal. Stone continues to play an important role in the MB declining significantly at first, but appears to recover in the MB IIC.

Recognition that stone tools continue to play a role in daily life activities, such as animal butchering, is critical as many Bronze Age excavations do not systematically collect the debris associated with stone tool production, use and discard. This leads to an emphasis upon the new technology at the expense of understanding the continued importance in different periods, subperiods and depositional contexts of the older technologies. It is not until the constituent parts of metal (tin and copper) become abundant enough (and thereby less expensive) that they become more accessible to most of the population and are eventually adopted for quotidian purposes.

As demonstrated here, there are many pathways for investigating the spread and rate of adoption of new technologies. This is particularly relevant for the Near East which is one of the earliest centres for the development of a metallurgical technology.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Lloyd Austin Courtenay, Université de Bordeaux, France

Reviewed by: Jerome Reynard, University of the Witwatersrand, South Africa

Verónica Estaca-Gómez, Complutense University of Madrid, Spain