Is It Worth It? A Review of Plant Residue Analysis on Knapped Lithic Artifacts

ABSTRACT

Residue analysis has become the go-to analysis for archaeologists in recent years. This review, focusing on the plant remains as residues from knapped lithic tools, has been carried out to understand the role of plant residues analyses in archaeological as well as associated experimental research, the advancements made in the field and the state of the art. Thirty-nine research papers starting from 1976 till 2020 have been analyzed critically in this review. In spite of numerous exemplary researches, it was observed that there are still many challenges to be overcome. The paper also makes a humble attempt at providing some suggestions for future research.

KEYWORDS:

• residue analysis
• knapped lithic tools
• phytolith analisis
• experimental research
• ethno-archaeological research
• plant remains

Introduction

Archaeology aims at understanding past societies and history of humankind through material remains. Although the aim of archaeology lies in understanding the processes in which the materials were involved, who were the social agents in charge of that and which was the social organization of each society, the study of objects is still the biggest protagonist within the discipline.

However, the fact that most materials used and consumed by prehistoric societies were perishable, and thus are difficult to be identified, does not allow for a proper reconstruction of past economies unless specific methodologies to understand them are developed (Hurcombe, 2008; Zurro, 2011). In this sense, archaeological methodologies and techniques have increased substantially during the last decades. The theoretical development of processual archaeology in the 1960s (McGovern et al., 1995; Watson et al., 1987) fostered a change in the archaeological discipline according to which achieving a high resolution in the identification of the materials became the center of archaeological research. That approach placed cultural chronologies, which had played a predominant role till this moment, in a secondary position and fostered the development of several archaeological areas of research such as archaeobotany or ictioarchaeology (to name a couple), which started to become standard at this moment and all of them eventually became essential during the 80s (Vicent, 1982).

Despite lithic being the most ubiquitous and in-depth analyzed archaeological remain, its study also went through changes during this period. The relationship between humans and lithic tools is an intimate one and it can be very well said that chipped stone tools and debitage represent the most abundant form of artifacts found from prehistoric sites. In many areas of the world they represent the only form of remains that have withstood the inroads of environmental and human perturbation, such as erosion, decay and landscape development. In fact, lithic artifacts represent one of the most important clues to understand prehistoric lifeways. As an archaeological artifact, they are the end result of a long process, which starts with the selection of raw material, designing and shaping, devising ways to haft, utilizing and ultimately discarding the tool. Thus, when trying to understand a tool, all these dynamic aspects need to be taken into consideration (Andrefsky, 2005; MacDonald and Wilkins, 2010; Moloney and Shott, 2003; Odell, 2004; Shott, 2014; Yerkes and Kardulias, 1993).

Stone tools have been considered crude markers of time and cultural affinity, have been subjects of study including the technological processes by which the tools were made as well as their implications for their makeŕs cognitive faculties (Shott, 2014). Tool function also became a big issue for lithic studies, and despite use-wear developed in the 50’s by Semenov (see Semenov, 1981 – or. 1957) it was also in the 80’s when it became popular. In fact, the first international conference was held in 1977 in Burnaby, Canada (Hayden & Kamminga, 1979). From the very beginning of the application of the use-wear method in Western Archaeology, there was a specific concern with understanding the appearance of the micropolish and its possible relationship with plant processing (Anderson, 1980; Mansur-Franchomme, 1983). During those researches, X-ray dispersion techniques applied in combination with scanning electron microscopy demonstrated the presence of particles of the worked material embedded in the micropolish layer. Apart from specific studies on micropolish, there was also a particular concern regarding residues preserved on tool edges in order to disentangle tool function (see Anderson, 1980; Anderson-Gerfaud, 1981, 1986; Mansur-Franchomme, 1983; Shafer & Hollowat, 1979, etc. amongst others). Thus, from the very beginning residues were part of the equation of addressing the use of a tool in which tool morphology, raw material and wear were also involved.

Residue analysis in Archaeology

Since the 80’s archaeology has reached a degree of sophistication never expected before. New techniques appeared in the specialized literature under different headings such as biomolecular archaeology, archaeometry or microfossil analysis to cite some of them. Residue analyses, one of these new avenues of development for our discipline, comprises a heterogeneous set of analyses also applied to lithic studies.

There is a big corpus of research labeled as “residue analyses” that analyzes floors, ceramic contents and remains in specific artifacts such as pipes or other unique objects or contexts (Barton, 2007; Brettell et al., 2017; Craig et al., 2020; Echeverría et al., 2014; Evershed, 2008; Heron et al., 2013; Heron & Evershed, 1993; Koh & Birney, 2019; Nugent, 2006; Pecci et al., 2013; Rosell et al., 2013; Sacchi et al., 2020; Tushingham et al., 2018, among others). These studies are mostly carried out through chemical and protein analyses and they usually correspond to late prehistory or to historical contexts.

Similarly, residue analyses are also carried out as part of forensic and burial archaeology. That is the case for the identification of different remains (such as plant fibers and phytoliths) in coprolites and stomach contents (Gilbert et al., 2008; Sutton & Reinhard, 1995). Phytoliths recovered from dental fissures (Ciochon et al., 1990) as well as in dental calculus are often referred to as residues too (Juan-Tresserras et al., 1997; Scott-Cummings & Magennis, 1997) and even though in these cases they are usually interpreted as food remains, they might also be understood as remains from chewing, biting or related with working processes (Oxenham et al., 2002; Radini et al., 2017).

Although most studies aim at identifying remains from the plant or animal realm (see Hendy, 2021 for a review on protein analysis), minerals used as pigments or as medical ingredients and even salt can be also part of analyzed residues (Guerra-Doce, 2015; Matarrese et al., 2011; Morell-Hart, 2022; Pérez-Arantegui, 2021). Archaeology has also explored stains, usually of extracted residues, that have the advantage of being an easy and rapid way to examine the presence of certain molecular structures (e.g. collagen) (Fullagar et al., 2015; Lamb & Loy, 2005; Rots et al., 2016; Stephenson, 2015).

In spite of residue analyses referring to a wide variety of case-studies, they share some common features;

• they aim at studying

  • a specific and sparse material which is generally not distinguishable to the naked eye (but whose presence is assumed)

  • the different ingredients of a composite given material

• they generally attempt to identify the function of an artifact or tool

• in most cases it becomes necessary to apply the latest technique for the visibilization and/or identification of the residue or combination of residues

• these studies mostly develop the technical side of the research (that which is based on the identification), rather than methodological as well as theoretical issues

Moreover, while some of these approaches rely mostly on reactive procedures (such as those made on pavements, resins, or blood residues), some others are based on the analyst´s capacity to carry out the visual morphological identification of the residue. That’s the case for the identification of microbotanical remains, which is one of the increasing procedures for the analyses of residues in lithics (Haslam, 2009).

In brief, the development of technical capabilities for material identification is allowing residue analysis to play a relevant role within archaeological studies, and it might become a standard for the discipline in the near future (Croft, 2021; Polla & Springer, 2022; Reber, 2022).

State-of-the-art and research objectives; challenges and achievements in lithic residue analysis

Residue analysis, started by Briuer (1976), has followed closely on the footsteps of microwear analysis, developed by Semenov and followed by many others, attempting to identify the materials the tool might have been directly worked on and which have left some “trace” behind to be identified.

In general, studies based on residue analyses aim at addressing different issues that range from understanding the artifacts themselves to more general questions at the landscape management level; understanding the micro-polish observed on the lithic tools (Anderson, 1980), and its relation with the presence of ashes (Lemorini et al., 2020) determining stone tool function (Fullagar, 1986; Kealhofer et al., 1999; Dominguez et al., 2001), identifying types of plants or animals being processed (Berman & Pearsall, 2020; Jahren et al., 1997; Hardy & Garuf, 1998) or identifying types of materials used to haft these tools (Sobolik, 1996). Other studies aim at hypothesis evaluation and innovative interpretations (Charrie-Duhart et al., 2013) and at the development of blind tests to understand research validity (Hardy & Garuf, 1998). Recently, new techniques such as scanning electron microscope (SEM) with energy dispersive X-ray spectroscopy (EDS), or chemical and elemental analysis such as Fourier transform infrared microspectroscopy (FTIRM), Raman microspectroscopy, X-ray fluorescence spectroscopy (XRF) and gas chromatography mass spectrometry (GC-MS) etc. are increasingly being incorporated into residue analysis as well (Croft, 2021).

In spite of the emphasis given to these analyses in recent years (e.g. Croft, 2021), what we can see is that residue studies performed on lithics are lacking not only standardized procedures but also accepted definitions. Although residue analysis as a field of study is mainly being defined due to the involved technical procedures, several theoretical and methodological issues, however, need to be considered.

First, it becomes critical to evaluate what we understand as residues. Residues can be remains of the refuse, byproducts that appeared during production and consumption processes and that are voluntarily discarded. Residues are also the involuntary remains left after a working process has taken place and they can also be a mixture of substances corresponding to single or multiple events. Although generally speaking, residues can be defined as any type of leftover or substance left behind; residues are substances that might correspond to (see Zurro, 2006 for an extended definition of the following):

• a raw material that has been processed

• a secondary material that has been processed

• an auxiliary material; that is, a material that helps in manufacturing particular resources, speeding up processes, improving results or achieving a specific finishing touch

• a material used for creating a composed tool, helping in assembling different pieces

Both the heterogeneity and lack of experience of these approaches performed on knapped lithics does not allow to assess to which extent they produce reliable as well as reproducible data, given the uniqueness of the samples and the impossibility in most cases to replicate the analysis. In addition, it is important to remark that given that very often residues are a combination of materials, the technique chosen for carrying out the analysis tends to drive the results achieved, thus eliminating possibilities to widen materials identification.

This review will give an overview of residue analyses applied to knapped lithics, aiming to identify general patterns regarding the specific objectives of the research and the specific laboratory and microscopy protocols followed by each research. Thus, our main objectives are as following:

• giving an overview of the state-of-the-art of residue analysis on knapped lithics

• find the variability in the procedures that are followed but also the shared approaches and/or protocols

• identifying the developments achieved until today

• identifying the still existing challenges

• assessing the role played by residue analyses in archaeological research

We will focus particularly on plant residues, as this is also a field of expertise that has grown exponentially during the last years and is part of the author’s expertise (Cabanes, 2020; Lancelotti & Madella, 2018; Rashid et al., 2019; Strömberg et al., 2018; Zurro et al. 2016) and that it is also part of the historical development of this specific area of research. As mentioned, earliest studies dealt with understanding the mechanisms of polish formation, a process that was somehow related with plant processing (Anderson, 1980; Fullagar, 1991).

Methodology

This paper is based on a literature review to understand the above mentioned objectives. The general methodology consisted in selecting publications representative of the topic under study, dissecting available information following some criteria a priori defined and following a straightforward analysis of the results.

Literature search was carried out in Scopus/WOK using the keywords “residue” and “lithic tools” appearing in the title as well as in abstracts, both in Spanish and English, as we consider they cover the biggest section of research carried out on this area of analysis. The sorting of literature was carried out between October 2020 and March 2021. These publications were narrowed down by selecting publications dealing with knapped lithic tools. The last stage involved further narrowing the search to those papers dealing with plant residues. Important references (which might have not been initially identified by Scopus/WOK) were also selected while carrying out the reading of the literature.

Ultimately our survey retrieved 39 documents, which constitute a sample covering a time span of 44 years starting with ¨New clues to stone tool function: plant and animal residues¨ (Briuer, 1976) until 2020 (Berman & Pearsall, 2020; Fuentes et al., 2020; Nucara et al., 2020).

Different information regarding research objectives, sample size, sampling procedures, methodological and technical aspects as well as criteria followed for the interpretation of the results were recorded in a new database (see Table 1 with the recorded information and the entire database as Supplementary Material).

Table 1. Recorded information from the literature survey with an example from Kealhofer et al., 1999.

Category	Variable	Example
Reference	Author(s)	Kealhofer L., Torrence R., Fullagar R.
	Title	Integrating Phytoliths within Use-Wear/Residue Studies of Stone Tools
	Journal	Journal of Archaeological Science 26
	Year of publication	1999
Research Goal	Type of material	archaeological
	Research objectives	investigate two methodological issues: 1. the integrity of phytolith residues and 2. interpretations of artifacts use based on phytolith assemblage composition and use-wear/residue analysis
Sample	Context	artifacts were selected from arbitrary unit of 1m x1m
	Tool type	flakes
	Raw material	quartzite, flint etc
	n (number of tools analyzed)	17
	n/x (tools selected for analysis/the total number of excavated tools)	not mentioned
Sample quality assessment	Control Sample (yes/no)	no
	Contamination considered (yes/no)	no
Method	Type of residue (plant, animal, etc.)	plants
	Overview of the methodology (in situ, extraction, use wear, combination)	Extraction use wear
	Extraction (scraping, water spraying, chemical spraying (Hcl, H2O2), brushing, ultrawave)	ultrawave
	Lithic tool sampled area (one edge, both edges, whole tool, not mentioned)	whole tool
	Type of microscopy employed (optical microscopy, SEM etc)	optical microscopy
Results	Qualitative/quantitative/both approaches	both
	Taxonomic results	predominance of bamboos, palms and arboreal taxa
	Interpretation	phytolith assemblages represent residues resulting from pre-depositional contact with materials. In case of obsidian artifacts-phytolith analysis makes a very important contribution to the understanding of chronological changes in tool use.
Observations	(By the reviewers)	pioneering and fundamental article to understand phytoliths and associated tool use

Results from the review

First of all, results show that there is a clear and progressive increase of articles related to residue analysis that runs in parallel to an increased use of techniques in archaeology together with the spread of multi-proxy approaches (Birks & Birks, 2006; Fuks & Dunseth, 2021; García-Granero et al., 2015) and an expanding period of phytolith and starch research (Figure 1; Hart, 2016 and a review in Rashid et al., 2019).

Figure 1. Graph bars showing the increase of publications regarding phytoliths and starches as retrieved through a SCOPUS search (search carried out on February 24th 2022). Regarding starch research, it also included the keyword “archaeology” as there are several researches within the food production research field dealing with starch.

Thirty-one out of the 39 documents fall within the period 2000–2020 (Figure 2), which can be understood as part of the tendency mentioned above. Regarding the journals, they are a heterogeneous assemblage of peer-reviewed publications with a special role played by Journal of Archaeological Science (11/39) and Journal of Archaeological Science Reports (4/39), together with Journal of Human Evolution (5/39) (Figure 3).

Figure 2. Graph showing the time span of the sample under study.

Figure 3. Graph showing the different journals in which the papers have been published.

In the following sections, we will review the different aspects regarding the various steps which should be involved in any research, comprising (a) General approach and research goals, (b) Materials, (c) Sampling and representativeness, (d) Methodological issues (laboratory and microscopy procedures) and, finally, (e) Results and interpretation.

a.	General approach and research goals: substantive versus methodological research

Research questions are directly related to the type of selected material. Most of the materials have been extracted from archaeological sites and museum collections (corresponding to 51% of the cases, Figure 4). While these case studies address substantive archaeological research questions (understanding specific issues related to the environmental or cultural context or the chronology under study), it was found that methodological approaches are obviously fundamentally carried out with experimental material, raising up to 33% of the recorded cases in our sample.

Figure 4. Graph showing the type of the analyzed materials. A_EXP shows studies where a combination of archaeological and experimental materials has been included.

While some studies use a combination of materials coming from different sources (Figure 4 shows ethnoarchaeological and combined approaches – A_EXP together form 16% of the cases), most of them focus on a particular type of record and archaeological together with experimental lithics represent 84% of the whole sample.

The review allows us to distinguish archaeological research goals that aim at creating specific data regarding past economies and technology, and those that explore different methodological issues (such as the viability or robustness of the methodology of residue analyses), both approaches corresponding clearly to archaeological and experimental studies respectively. However, residue analyses have been carried out to answer more questions that a priori might be considered (Table 2).

Table 2. Identified research topics.

Archaeological questions			Methodological issues
Plant domain	Tool use	Spatial analysis	Residue	Sample quality	Integration
Environment	Efficiency (retouching)	Activity areas	Identify	Preservation	Use-wear
Plant management	Study of specific tools		Recover	Contamination	Polish
Plant consumption	Cinematics, Polish formation		Identify taxa
Food consumption	Handling

Since we have decided to focus on plants, the most obvious question addressed by researchers is identifying working processes in which plant material was involved (Jahren et al., 1997; Kealhofer et al., 1999; Rots & Williamson, 2004). Preservation of plant residues on tools is a major question addressed by researchers (Langejans, 2010; Wadley et al., 2004), but also understanding plant management or consumption (Aceituno et al., 2001; Berman and Pearsall, 2008, 2020; Colobig et al., 2020; Fuentes et al., 2020; Hardy et al., 2018; Hardy & Moncel, 2011; Henry et al., 2014; Piperno et al., 2009). This goal can be achieved at a very general level (landscape management and anthropisation of natural environments; e.g. starting agricultural practices or harvesting of particular species) or at a lower scale, identifying the types of plants or even specific taxa consumed as part of the diet (Fullagar et al., 2006; Liu et al., 2018).

Part of the papers also deal with the function of the tools themselves, so that residue analysis is carried out to complement lithic methodologies or to discuss use-wear results (Aleksandrova et al., 2014; Cueto et al., 2010; Hardy et al., 2001; Unger-Hamilton, 1984). Therefore, specificities of the artifacts employed are also paramount, trying to deepen into particular lithic types (Domínguez Rodrigo et al., 2001; Hardy et al., 2018; Liu et al., 2017; Venditti et al., 2019), and the study of retouching efficiency (Hardy & Garuf, 1998; Lombard, 2005), employed cinematics or handling (Sobolik, 1996). Finally, other archaeological research aims to spatially integrate residue analyses and lithics, which might add information to identifying activity areas (Álvarez et al., 2009; Ochoa et al., 2013).

Within the methodological approach, studies have also been conducted to acknowledge different types of problems, mostly dealing with identification of specific residues or contaminations of residue analysis (Monnier et al., 2012; Xhauflair et al., 2017). In this group there are particular studies related to creating reference collections or how to carry out interpretations based on controlled experiments (Lynch & Miotti, 2017; Nucara et al., 2020; Pedergnana & Olle, 2018). Regarding methodological issues, these range from recognizing various types of residues, their recovery and identification, assessing sample quality, understanding which mechanisms allow preservation, and which dynamics might produce sample contamination. Finally, it is paramount to understand how residue analyses can be made compatible with use wear research (specific working processes) and how results coming from both analyses can be integrated or help solve specific research questions from any of the areas (such as explaining polishing as a general feature and its relationship with silica).

b.

Materials

Scholars have analyzed almost all types of stone tools, ranging from simple flakes to sophisticatedly made tools. In-depth analysis showed that 41% (n = 16) of the total papers have flakes as their main tool types, specifically six papers in the archaeological category, eight in the experimental category and two papers which dealt with combined materials, archaeological as well as experimental. Many of the papers had also analyzed other forms of tools along with flakes, e.g. microliths, grinding stones, choppers and chopping tools. Other tools to receive special focus are points, heavy duty tools, scrapers and denticules. Four papers have focused on broader issues and for this, have analyzed various forms of tools altogether (Table 3).

Table 3. Types of tools analyzed by the scholars.

N. of papers	Tool type	References
16	Flakes	Berman and Pearsall, 2000; Briuer, 1976; Colobig et al., 2020; Fuentes et al., 2020; Fullagar, 1991; Hardy & Garuf, 1998; Jahren et al., 1997; Kealhofer et al., 1999; Langejans, 2010; Ochoa et al., 2013; Lynch & Miotti, 2017; Nucara et al., 2020; Pedergnana & Olle, 2018; Piperno et al., 2009; Venditti et al., 2019; Wadley et al., 2004
04	Microliths	Anderson, 1980; Aleksandrova et al., 2014; Hardy et al., 2001; Unger-Hamilton, 1984
04	Scrapers	Debert & Sherriff, 2007; Monnier et al., 2012; Rots & Williamson, 2004; Sobolik, 1996
04	Various types of tools	Álvarez et al., 2009; Berman & Pearsall, 2020; Hardy et al., 2008; Henry et al., 2014
02	Flakes and microliths	Domínguez Rodrigo et al., 2001; Hardy et al., 2018
03	Flakes and grinding stones	Fullagar et al., 2006; Liu et al., 2017, 2018
01	Point	Aceituno et al., 2001
01	Heavy duty tools	Lombard, 2005
01	Scrapers and points	Hardy & Moncel, 2011
01	Flakes, choppers and chopping tools	Xhauflair et al., 2017
01	Scrapers and denticulates	Cueto et al., 2010
01	No informatio	Pearsall, Duncan, Chandler-Ezell, Ubelaker, & Zeidler, 2020

The fact that the majority of the papers have taken flakes as their main tool types for conducting residue analyses (Table 3) is not surprising since flakes are one of the most ubiquitous artifacts belonging to a lithic assemblage. It has also been demonstrated that macroscopically unmodified artifacts have been used almost as intensively as modified artifacts (Vaughan, 1985) which makes flakes ideal tool types to be selected by scholars.

Analysis of raw material showed that most of these tools (total papers n = 16) were made out of crypto crystalline materials such as flint, chert etc. However, non crypto crystalline materials have also been analyzed (total papers n = 9). Five papers dealt with both types of materials and ten papers have not given information about the raw material (Table 4). While archaeological examples show a heterogeneous selection of raw materials, experimental studies, on the other hand, preferred to make their stone tools out of crypto crystalline siliceous stones (which are easier to knap).

Table 4. Overview of raw materials.

No. of papers	Raw material	References
16	Crypto crystalline	Anderson, 1980; Berman and Pearsall, 2000; Cueto et al., 2010; Fullagar, 1991; Fullagar et al., 2006; Fuentes et al., 2020; Hardy & Garuf, 1998; Hardy & Moncel, 2011; Jahren et al., 1997; Monnier et al., 2012; Nucara et al., 2020; Rots & Williamson, 2004; Sobolik, 1996; Unger-Hamilton, 1984; Venditti et al., 2019; Xhauflair et al., 2017
09	Non crypto crystalline	Álvarez et al., 2009; Colobig et al. 2020; Debert & Sherriff, 2007; Domínguez Rodrigo et al., 2001; Hardy et al., 2001, 2018; Kealhofer et al., 1999; Lombard, 2005; Lynch & Miotti, 2017
05	Both types	Berman & Pearsall, 2020; Langejans, 2010; Liu et al., 2017; Pedergnana & Olle, 2018; Wadley et al., 2004
09	No information	Aceituno et al., 2001; Aleksandrova et al., 2014; Briuer, 1976; Hardy et al., 2008; Henry et al., 2014; Liu et al., 2018; Ochoa et al., 2013; Pearsall et al., 2020; Piperno et al., 2009

The identification of use-wear traces is paramount to understand what part of the tool was the active one. Therefore, when it comes to the question of what part of the tool was sampled and/or analyzed, it can be seen that archaeological as well experimental papers have most of the time (n = 20) focused on the active edges of the tools. However, one paper has analyzed the haft area in addition to the active edges. Altogether, fourteen papers have examined either the whole tool, various parts of an edge, e.g. the tip, mid-section etc. or have considered ventral/dorsal sides of a tool separately. Remaining papers (n = 4) have not made this information available (Table 5).

c.

Sampling and representativeness Table 5. Consideration of the tool edge. No. of papers Wear traces Reference 20 Active edge Anderson, 1980; Álvarez et al., 2009; Berman & Pearsall, 2020; Briuer, 1976; Colobig et al. 2020; Fullagar, 1991; Hardy & Garuf, 1998; Hardy & Moncel, 2011; Henry et al., 2014; Jahren et al., 1997; Langejans, 2010; Liu et al., 2017, 2018; Lynch & Miotti, 2017; Ochoa et al., 2013; Pedergnana & Olle, 2018; Piperno et al., 2009; Unger-Hamilton, 1984; Wadley et al., 2004; Xhauflair et al., 2017 14 Whole tool Aleksandrova et al., 2014; Debert & Sherriff, 2007; Domínguez Rodrigo et al., 2001; Fuentes et al., 2020; Fullagar et al., 2006; Hardy et al., 2001, 2008, 2018; Kealhofer et al., 1999; Lombard, 2005; Monnier et al., 2012; Nucara et al., 2020; Rots & Williamson, 2004; Venditti et al., 2019 01 Haft area + active edge Sobolik, 1996 04 No information Aceituno et al., 2001; Berman and Pearsall, 2000; Cueto et al., 2010; Colobig et al., 2020; Pearsall et al., 2020

Sampling

Context is one of the most important aspects of archaeology. It gives artifacts their legal authenticity and archaeological significance (Ford, 1977). Archaeological papers (n = 10) describe the layer from which the analyzed artifacts were recovered (Table 6). Eight papers give general information about the site but without giving further details regarding the corresponding context. The archaeological layer was described in the combined papers (archaeological and experimental) four times. The rest of the archaeological papers have not given details regarding this aspect (Table 6).

Table 6. Details of the context (13 papers have not been considered as they deal with experimental, ad hoc produced material).

No. of papers	Context	References
10	Layer	Aceituno et al., 2001; Aleksandrova et al., 2014; Álvarez et al., 2009; Dominguez et al., 2001; Fullagar et al., 2006; Hardy & Moncel, 2011; Kealhofer et al., 1999; Ochoa et al., 2013; Pearsall et al., 2020; Sobolik 1996
08	General information about the site	Berman and Pearsall, 2000, 2020; Briuer, 1976; Debert and Sheriff 2007; Hardy et al., 2008, 2018; Henry et al., 2014; Piperno et al., 2009
04	Layer (combined papers)	Fuentes et al., 2020; Lombard, 2005; Venditti et al., 2019; Rots and Williams 2004
04	No information	Anderson, 1980; Colobig et al. 2020; Hardy et al., 2001; Liu et al. 2018

Sample size

Understanding the sample size and assessing the strength and reliability of the results was one of the biggest challenges faced during this research. All the experimental papers clearly mention the number of artifacts analyzed. These range from papers giving results of experiments with two artifacts to 99 artifacts. The papers analyzing both archaeological as well as experimental materials, do give information about the number of experimental artifacts but fail to mention the total archaeological assemblage size.

Regarding archaeological research, in all cases we have the number of artifacts comprising the set under study but not necessarily the total amount of lithics found in the corresponding archaeological context. Thus, the representativeness of the sample could not be assessed correctly (Table 7).

Table 7. Sample size per each research, showing the number or artifacts analyzed and the lithic assemblage, also showing not available data (N/A). Label A corresponds to archaeological research, EA to ethnoarchaeological, EXP to experimental and A_EXP to those studies combining archaeological and experimental.

Authors	material	n  (artefactsanalysed)	n  (total lithicassemblage)	%
Aceituno et al. 2001	A	1	N/A	-
Aleksandrova et al. 2014	A	153	3,700	4.14
Alvarez et al. 2009	A	33	N/A	-
Berman & Pearsall 2020	A	25	N/A	-
Berman & Pearsall 2000	A	6	N/A	-
Briuer 1976	A	37	2,551	1.45
Colobig et al. 2020	A	1	6	16.6
Debert & Sheriff 2007	A	26	3,000	0.87
Dominguez-Rodrigo et al.2001	A	5	N/A	-
Fullagar et al. 2006	A	12	N/A	-
Hardy & Moncel 2011	A	182	N/A	-
Hardy et al. 2001	A	50	N/A
Hardy et al. 2008	A	109	N/A	-
Hardy et al. 2018	A	99	N/A	-
Henry et al. 2014	A	209	N/A	-
Kealhofer et al. 1998	A	17	N/A	-
Ochoa et al. 2013	A	8	N/A	-
Pearsall et al. 2020	A	50	N/A	-
Piperno et al. 2009	A	5	N/A	-
Sobolik 1996	A	55	N/A
Rots & Williamson 2004	EA	18A+ 20E	N/A	-
Anderson 1980	A_EXP	13A+ 23E	N/A	-
Fuentes et al. 2020	A_EXP	51A+ EXP(N/A)	14, 525	1.25
Liu et al. 2018	A_EXP	26A+ EXP(N/A)	N/A	-
Lombard 2005	A_EXP	50A+ 35E	N/A	-
Venditti et al. 2019	A_EXP	609+ EXP(N/A)	N/A	-
Cueto et al. 2010	EXP	12	12	100
Fullagar 1991	EXP	81	81	100
Hardy & Garufi 1998	EXP	N/A	N/A	-
Jahren et al. 1997	EXP	3	3	100
Langejans 2010	EXP	160	160	100
Liu et al. 2017	EXP	17	17	100
Lynch & Miotti 2017	EXP	26	26	100
Monnier et al. 2012	EXP	18	18	100
Nucara et al. 2020	EXP	70	70	100
Pedergnana & Olle 2018	EXP	23	23	100
Unger-Hamilton 1984	EXP	2	2	100
Wadley et al. 2004	EXP	28	28	100
Xhauflair et al. 2017	EXP	99	99	100

Out of 20 papers analyzing archaeological material, only five (25%) supplied information regarding the total number of artifacts from their archaeological assemblage. Briuer (1976) mentions the total number of the artifact assemblage (n = 2551) but it has not been made clear how many, if all, actually underwent the analyses. Debert and Sherriff (2007) analyzed 70 (2.33%) out of 3000 formal tools, and 26 (out of the already selected 70) for SEM (making a final sample of 0.86%). Aleksandrova et al. in 2014 analyzed 765 (20%) artifacts out of the 3700 for use-wear analysis though not giving precise information regarding residue analysis. Fuentes et al. (2020) mention that from a total of 14,525 lithic artifacts, 183 (1.25%) were analyzed for usewear, further mentioning that 51 archaeological tools were identified with plant remains. For Fuentes et al. (2020) the total number, however, includes tools as well as lithic debitage and thus the analyzed tools might in fact represent a much smaller percentage.

d.	Methodological issues (laboratory and microscopy procedures)

The methodology (for the study of residues, e.g. extraction, microscopy, etc.) designed and followed has depended mostly on their respective research questions. The general (archaeological or experimental) approach has conditioned the expected results (see Figure 5). While for experimental case studies there is a wide variety of residues being identified, in archaeological cases these are generally narrowed given the need to eliminate organic matter to create better observation conditions (so that some of these residues are destroyed). Over the years, attempts have been made to refine earlier methodologies for better understanding the tool functions as well as incorporation of newer techniques for better results, producing a high variability of how the residues are studied.

Figure 5. Expected plant residues

Figure 6. Scheme showing a summary of protocols for studying residues in lithics in situ or through extraction.

Since the 80’s, identifying microwear patterns on the tool edges to ascertain tool use has become a standard practice among researchers working with archaeological artifacts. The edge wear analysis is mostly conducted utilizing an optical microscope, though there are instances when scanning electron microscope (SEM) has been utilized (Berman and Pearsall, 2000; Debert & Sherriff, 2007; Hardy et al., 2018). SEM was found to have been utilized more for studies which incorporate analysis of both archaeological as well as experimental artifacts (Anderson, 1980; Fuentes et al., 2020; Venditti et al., 2019). One study has used FTIR in combination with optical microscopy and X-ray fluorescence analysis (Aleksandrova et al., 2014) for archaeological samples.

During the 90s a change was seen in the methodology when the archaeologists started extracting the residues from the active edges of the artifacts for a better identification. By looking at the papers dealing with archaeological artifacts it can be seen that the extraction is generally carried out by a sonic cleaner (used with distilled water), however there are instances when the extraction was seen to have been conducted by laboratory picks, pipettes or even by dry brushing of the artifacts (Figure 6, Table 8). Researchers have not followed a standard time for the ultrasonic extraction and whenever mentioned, it can be seen that the sonication time varies between 5 min (Kealhofer et al., 1999) to 10 min (Venditti et al., 2019). It was further observed that no standard way to mount the slides after the extraction has been followed. Apart from majority of the papers not mentioning the exact step-by-step procedure, mounting of the slide varied from using glycerin (Sobolik, 1996), cinnamaldehyde (Domínguez Rodrigo et al., 2001), Karo^TM and benzyl benzoate (Fullagar et al., 2006) to distilled water and glycerol (Pearsall et al., 2020). Further, only three papers clearly mention either the number of tracks on the slides counted for phytolith identification (Pearsall et al. 2020) or actual phytoliths counted per slide (Álvarez et al., 2009; Kealhofer et al., 1999). For the identification and naming of the residues, apart from not giving any information on this, scholars have either taken help from their own reference collection (Briuer, 1976; Fullagar et al., 2006; Hardy et al., 2001; Kealhofer et al., 1999; Piperno et al., 2009), previously published references (Debert & Sherriff, 2007; Domínguez Rodrigo et al., 2001; Hardy et al., 2001, 2008, 2018; Hardy & Moncel, 2011; Sobolik, 1996) databases such as University of Missouri Phytolith database or the International Code for Phytolith Nomenclature 2005 (Berman & Pearsall, 2020; Henry et al., 2014; Pearsall et al., 2020).

Table 8. Various extraction methodologies.

No. of papers	Extraction method	Reference
10	Sonic cleaner	Aceituno et al., 2001; Álvarez et al., 2009; Domínguez Rodrigo et al., 2001; Debert & Sherriff, 2007; Fullagar et al., 2006; Kealhofer et al., 1999; Ochoa et al., 2013; Pearsall et al., 2020; Piperno et al., 2009; Sobolik, 1996
01	Laboratory picks	Domínguez Rodrigo et al., 2001
04	Dry brushing	Aceituno et al., 2001; Aleksandrova et al., 2014; Berman & Pearsall, 2020; Sobolik, 1996

Researchers dealing with experimental artifacts have most of the time followed a combined methodology of.

1. in situ identification with edge-wear analysis (Fullagar, 1991; Hardy & Garuf, 1998; Langejans, 2010; Unger-Hamilton, 1984; Wadley et al., 2004; Xhauflair et al., 2017);

2. extraction with edge-wear analysis (Fullagar, 1991; Jahren et al., 1997; Liu et al., 2017) or

3. optical microscopy in combination with SEM (Jahren et al., 1997; Lynch & Miotti, 2017; Monnier et al., 2012; Pedergnana & Olle, 2018; Unger-Hamilton, 1984). Most of them have utilized an optical microscope for their edge-wear identification (Cueto et al., 2010; Fullagar, 1991; Hardy & Garuf, 1998; Langejans, 2010; Liu et al., 2017; Lynch & Miotti, 2017; Monnier et al., 2012; Pedergnana & Olle, 2018; Wadley et al., 2004; Xhauflair et al., 2017).

It was also seen that SEM has been utilized more in studies dealing with experimental materials (Jahren et al., 1997; Lynch & Miotti, 2017; Monnier et al., 2012; Pedergnana & Olle, 2018; Unger-Hamilton, 1984; Xhauflair et al., 2017). One study (Nucara et al., 2020) has used FTIR along with Principal Component Analysis for their methodology to understand the potential of applying the PCA to discriminate different classes of vegetal species between a large number of FTIR spectra.

e.	Results and interpretation

Concerning the results, they show different degrees of detail regarding identifying the residues, ranging from very general results to those referring to a specific taxon. Some papers verify that the information obtained through use-wear can be confirmed through residue analysis (Hardy et al., 2001) or that specific cinematics have been used (Debert & Sherriff, 2007) corresponding to a particular working process (Fullagar et al., 2006). Following this trend, other researchers mention that residues are found in a specific area of the lithics. While Lombard (2005) relates in her research these results to hafting, in other cases there is no in-depth analysis of the implications regarding the methodology employed or how this contributes to the understanding of tool use.

Papers whose research addresses plant management or consumption offer quantitative and qualitative data. Qualitative data range from general information (referring to woody or starchy plants, see Fullagar et al., 2006 as an example) to increased accuracy in the taxonomic identification. In this sense, some publications speak about arboreal taxa (Kealhofer et al., 1999), grass (Domínguez Rodrigo et al., 2001) or palm phytoliths (Kealhofer et al., 1999); others refer specifically to maize or Marantaceae phytoliths among others (Pearsall et al., 2020). Many of these researches combine phytolith and starch analysis information, building a stronger archaeobotanical statement (Piperno et al., 2009), in which the potentiality for each technique is assessed (see Berman & Pearsall, 2020, where starches are underlined as giving much more information than phytoliths).

Regarding microbotanical remains analyses (phytolith and starches) it is essential to remark that only part of the papers gives quantitative information (number of phytoliths and starches identified neither in general nor per artifact). Álvarez et al., 2009, Liu et al., 2018, and Berman & Pearsall, 2020 were the only four papers to offer quantitative data. Within the specialized literature, phytolith data is usually provided as tables. This standard is followed in some articles (Álvarez et al., 2009), while others offer this information within the text. In addition, only part of them offer pieces of information about the diagnostic anatomical features used for identifying the phytoliths (Briuer, 1976; Fuentes et al., 2020; Sobolik, 1996) or refer to specialized phytolith analysis literature (Fuentes et al., 2020).

There are also studies integrating different types of residues. Paper by Aleksandrova et al., 2014 gives results obtained in a comprehensive analysis of organic and mineral residues on the stone tool surfaces which helped to reconstruct the ways of stone tool hafting, restore partial composition of glues and also provided additional information supporting use-wear analysis.

Regarding the interpretation of the results, some of the papers simply reinforce the validity of the methodology applied (Rots & Williamson, 2004) or connect the data produced through residue analysis with already existing available information (Piperno et al., 2009). Some others aim at underlaying how the results contribute to knowledge production in archaeology and how residue analyses enrich the archaeological record, either contributing to the lithics field of research discussing hafting (Aleksandrova et al., 2014) or the multipurpose use of the tools (Sobolik, 1996) among other topics, but also creating new till the moment unavailable data when it comes to plant processing (Domínguez Rodrigo et al., 2001; Hardy et al., 2001). The contribution of these new data to plant consumption either at site or regional scale and concerning specific chronological frameworks is assessed (Berman & Pearsall, 2020; Fuentes et al., 2020; Hardy et al., 2008, 2018; Hardy & Moncel, 2011; Henry et al. 2014; Liu et al., 2018).

Results from the experimental papers vary regarding identifying understanding the role of residue to form microwear polish (Anderson, 1980; Fullagar, 1991; Unger-Hamilton, 1984); as evidence of hafting (Lombard, 2005); understanding specific types of residues, e.g. wood (Hardy & Garuf, 1998), plants (Colobig et al., 2021), bone and bamboo (Jahren et al., 1997); understanding residue preservation mechanisms (Langejans, 2010) and accuracy of residue identification (Wadley et al., 2004). Experiments have also been conducted to understand bigger issues such as the transition from wild to domesticated cereal (Liu et al., 2017) and general subsistence behavior during specific chronological periods (Aceituno et al., 2001; Fuentes et al., 2020; Ochoa et al., 2013; Venditti et al., 2019). Accuracy of residue identification (Wadley et al., 2004) and how to enhance this through new techniques and methods (Nucara et al., 2020) is another important area addressed by scholars. There are papers pointing out cautions to be followed by residue analysts (Lin et al., 2018; Marreiros et al., 2020; Monnier et al., 2012; Xhauflair et al., 2017). Scholars have also made reference collections to be used in future analyses (Lynch & Miotti, 2017; Pedergnana & Olle, 2018).

Discussion

Over the years, research on residue analysis has gained momentum incorporating newer and refined technologies. Research carried out till now has analyzed various types of stone tools belonging to various chronological periods and contexts to get a broader understanding of their form as well as function, producing a very heterogeneous set of methodologies and results.

As this review shows, the archaeological and the experimental papers have approached the field very differently as they correspond to two different modes of research design, dealing with different research questions and specific types of selected materials. Scholars working with archaeological artifacts have mainly tried to understand ancient plant consumption patterns by identifying the residues and understanding the function of lithic tools, while scholars working with experimental artifacts have mainly tried to refine the techniques as well as to incorporate newer technologies for better accuracy of the interpretation. At the same time, they produce different types of archaeological knowledge, generating databases for future references.

The design of the experimental studies allows the researchers to have almost absolute control over the different variables involved, allowing them to take conscious decisions. In fact, in most cases focus is on narrow or technical aspects of residue analysis such as understanding cinematics of tool use or possible contamination processes. Because of this, experimental research does not raise the question about the nature of the residues, but on the capability of the analyst to distinguish it and produce a general understanding of how they look under the microscope. The whole experiment is planned to be carried out in a short time, so that the lithics are used and analyzed without undergoing not only the archaeological post-depositional processes, but also storage processes that might result in contamination or loss of part of the residues. In addition, lack of sediment allows for a better identification of the residues, which is not the case for archaeological case studies.

In fact, the biggest difference between both types of research is that, while residues from archaeological materials are analyzed after they have been extracted from the tools, experimental studies tend to analyze them in situ. Due to this, a high variability is evident between researchers working with archaeological artifacts versus experimental material. On top of this, uniqueness of archaeological residues fosters their extraction from the lithics, so that they can be preserved mounted on slides with permanent media. Besides increasing their preservability, contamination processes that might take place in the laboratory are avoided.

Archaeological residues are not only unique, but also usually scarce and many times embedded within a mixture comprising different types of organic compounds ranging from collagen to soft tissues or oils, among others (see Regert, 2004 or Wadley et al., 2019) that makes it difficult to isolate a specific type of residue. In addition, archaeological deposits and sediments also contain variable organic matter richness that might transfer to the artifacts (Frahm et al., 2022). Thus, caution must be followed when selecting a specific technique and, therefore, a specific procedure for the study of the residues, as many times it conditions preservation of other types of remains (see Newman & Julig, 1989; Kealhofer et al., 1999; Hardy et al., 1998; Lemorini et al., 2019) (e.g. while removing organic matter we are destroying starch grains). In spite of the differences, it was observed that the procedures for the study of residues, e.g. extraction, microscopy etc. designed and followed by the researchers, mostly depended on their respective research questions.

Regarding the developments achieved until today, similar to archaeology, lithic analysis has proven to be an ever-evolving field which keeps incorporating newer techniques and methods. Inclusion of refined techniques, e.g. SEM or FTIR in the field of residue analysis has helped the field to grow continuously. Other than applying newer techniques and being able to identify (reaching different layers of taxonomic accuracy) the animal, plant or mineral origin of the residues, researchers have also taken help from statistical methods such as Principal Component Analysis (Nucara et al., 2020) to validate their results. In spite of all the achievements mentioned above, there are many challenges still to be overcome in this field.

To start, it becomes difficult to assess the state-of-the-art of residue analyses in lithics because of the diverse nature of the artifacts under study. Most of the time, the criteria followed for selecting the sample at both qualitative and quantitative levels is not always clear. In theory, the sample should represent a target population. Despite seemingly obvious, this was one of the most difficult parameters to analyze for us. Most papers deal with more than one tool type, sometimes coming from more than one layer, or from different areas of excavation, corresponding to different chronologies, so that it becomes difficult to assess whether they would like to understand the lithic assemblage (lithic remains recovered from a single archaeological context; e.g. area, layer, site, region (Hardy et al., 2001); or a lithic toolkit (a specific tool type (Debert & Sherriff, 2007), or a specific set of tool types belonging to a known archaeological period( Fullagar et al., 2006).

This creates a theoretical problem in which the sample quantitatively but also qualitatively might not be representative of the context under study or of the research question. In fact, in some of the research that has been reviewed, sample size corresponds to a minute piece of evidence, which does not allow for a proper assessment of the behavior of involved variables. Many times, it was found that papers had overlooked mentioning their tool type as well as the total sample size, creating a general discrepancy in numbers, leading to confusion regarding a general comparison or understanding.

Most papers show a lack of detailed information regarding methodology which does not allow for a replication of the study with respect to laboratory procedures. This has led to scarcity of robust results which can be used for understanding patterns of behavior of all the variables involved for addressing new or future research.

Regarding quantification of results, this is probably the most difficult variable to analyze. How do we decide how much residue has to be present on the edge to confirm it is a remain of a performed task? More research needs to be carried out before being able to answer this question properly. Although residue analyses can be developed independently, residues on stone tools are key for understanding use, and can be used as an anchor for developing residue analyses methodologies.

The biggest issue that we have encountered is the problem of integration of both types of research; experimental and archaeological. Experimental research is generating basic knowledge under controlled conditions with ad hoc manufactured lithics. It aims at becoming a reference point for precise issues such as recording how particular residues look under the microscope, on the tool edges themselves or for the associated micropolish. While this is true, it often easily overlooks several important variables such as the effects of post depositional conditions or decay. At the same time, experiments are conducted under strict laboratory conditions which are far from being representative of the real use of ancient tools. In spite of experimental research producing valuable insights for residue identification and understanding of preservation conditions, it seems that outcomes from experimental research are falling somehow short in many ways to lead archaeological queries. Thus, experimental and archaeological residue research are becoming two separate fields of knowledge that could benefit further from each other. Some of the analyzed papers do deal with both types of materials, however, it was observed that research questions at the experimental and archaeological level need to share an equal scale of analysis.

In addition, there is a need to work further in the relationship between the processing of different plant organs/tissues/species and the production of specific types of use-wear and remaining of residues on the artifacts through extensive experimentation (see Lemorini et al., 2014) as it has been carried out with animal tissues (see Monnier and May, 2019).

Last but not least, the backgrounds of the authors, who are mainly lithic specialists has conditioned their research questions, which are many times related to widening the understanding of the precise utilization aspects of these tools, while methodological precautions coming from other specialized fields, such as the phytolith field, might be inadvertently overlooked.

According to our standpoint, there are certain good practices that might standardize the production in this field of knowledge and increase the likelihood that researchers will be able to evaluate the findings of other researchers and the results better be comparable;

1. Research goals and objectives as well as the archaeological unit (site, deposit, layer, habitational unit) that will be analyzed and the adequateness of analyzing that unit for accomplishing the objectives.

2. Description of the study sample.

   • n of analysed as well as n of total artifacts per study unit.

   • criteria (size, appearance, type, …) followed for selecting the sample.

   • sampling strategy (systematic, random, x number of artifacts per archaeological unit, etc.).

   • location of the samples.

   • description of the sedimentological deposit.

3. Description of tools types and raw materials.

4. Absence/presence of control sample and description of it.

5. Contamination, if any, has been considered.

6. Method for analyzing the residues; comprising a detailed description of laboratory methodology, chemicals used and sampled tool area.

7. Type of microscopy employed, counting and identification protocol.

8. Use of reference collection, if any.

Conclusions

Over the years, residue analysis has provided detailed and compelling evidence of specific food, technological activities as well as tool manufacturing choices (Croft, 2021) and will continue to provide exemplary results to understand both the tools as well as the people behind the tools whenever proper procedures and precautions are followed. To address the question whether is it worth it, we think that residue analysis absolutely is fundamental though there still are several aspects of this field of research that need to be further developed to be more informative for our discipline. As a field of study, residue analyses is a paradigmatic example of the development of archaeology as a discipline; there is a high degree of technification addressing the identification of materials which is not always followed by a corresponding archaeological interpretative framework. This needs to be, for sure, one of the future avenues of development of residue analyses.

Acknowledgements

The authors are thankful to all the scholars whose studies have been included in this paper. We are grateful for their hard work and dedication to always evolve the field of archaeology and our understanding about our ancestors and their lifestyles. This paper has no intention of undermining any research, but to characterize the state-of-the-art of the topic. We are also thankful to the anonymous reviewers whose comments have enriched our paper. The publication fee was supported by the CSIC Open Access Publication Support Initiative through its Unit of Information Resources for Research (URICI).

Disclosure statement

No potential conflict of interest was reported by the authors.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work has been supported by MSCA -Individual Fellowships (Horizon 2020 Research and Innovation Programme, PAST - Phytolith Analysis and Stone Tools: A socio-ecological analysis of stone tools assemblages of North- Western South Asia, grant number 891238).