Bone tools have a long archaeological history, and have recently been shown to retain use-traces distinctive of different perishable crafting practices. When examined in a controlled way, these diagnostic use-traces can serve as proxies for the crafted forms the bone tools were used to produce (e.g., baskets, leather goods, etc.). However, a number of methodological stumbling-blocks have hindered the sharing of bone tool use-wear results in a consistent standardized format. We suggest the application of Reflectance Transformation Imaging (RTI), and provide details for how to construct an RTI system which resolves most problems related to reproducibility in use-trace analysis.
In archaeology, any understanding of the past is based on interpreting material remains left by past peoples. We do, however, realize that not all objects and technologies used in the remote past will have survived to the present day. In particular, items made out of organic materials, including crafted forms, are vastly underrepresented at archaeological sites (Croes 1997, Hurcombe 2008, Soffer et al. 2001). In most cases the archaeological invisibility of these crafted forms pairs with the archaeological invisibility of their producers to reveal a vast blind spot in our knowledge of how the remote past was actually lived, and by whom (Bradfield 2014, Griffits 2001, Langley et al. 2021, Medina et al. 2018, Stammers et al. 2017, Stone 2009, 2013).
Of the organic materials that do survive, bone is certainly the most common. Bone tools have been used by hominin species for millions of years, and are in evidence in a number of world contexts at least since the Lower Palaeolithic (e.g. Backwell and D’Errico 2001, Biddittu and Celletti 2001, Bouzouggar et al. 2019, Brain 1993, Mania 1962, Tourloukis et al. 2019, Yellen et al. 1995, Yizraeli 1967, etc.). Because the field of prehistoric archaeology has been historically dominated by the study of stone tools, the lesser attention given to bone tools has obscured understanding of their place(s) in the technological repertoire of Palaeolithic human groups (Antonites et al. 2016, Desmond et al. 2018, Soffer 2004, Stone 2009). Stone tools are virtually indestructible, and their tendency to survive has led to an over-emphasis on particular practices such as hunting and butchery. This, in turn, can distort modern researchers’ views of the scope of possible activities within prehistoric groups (Soffer 2004, Soffer et al. 2001, Stone 2009, 2013). Conversely, a lack of evidence for perishable technologies in turn elides the contributions and life histories of anyone not engaged in hunting: in other words, contributions of women (e.g., pregnant or nursing mothers), small children, the elderly, people with injuries, physical and mental disabilities, etc. (Soffer 2004, Soffer et al. 2001, Stone 2009, 2013). As with bone tools, the presence of crafted products is implied very early in our evolutionary history. For example, baby carriers are hypothesized to have been in use for millions of years, since Homo erectus (Berecz et al. 2020, Ehrenberg 1989, Falk 2009) if not earlier (Caldwell 2013, Taylor 2010). Though crafted forms themselves may not survive, in many cases bone tools used in craft production may still exist. It is only recently that, by combining tribological principles (that is, the science of interacting surfaces in motion) with experimentation, archaeologists have been able to distinguish bone tool use-wear characteristic of different crafting processes. As such, bone tools can serve as a crucial proxy for otherwise archaeologically invisible activities (Bradfield 2020, Desmond 2018, 2019, Langley et al. 2021, Soffer 2004, Soffer et al. 2001, Stone 2013).
In recent years, there has been a widespread recognition that morphological and/or typo-functional descriptions alone are not trustworthy indicators of how a bone tool was actually used (many of which can appear formally similar), and that the study of microtopographic wear traces apparent on the tools themselves is the only reliable way to diagnose use (Akhmetgaleeva 2017, Antonites et al. 2016, Arrighi et al. 2016, Bradfield 2012, 2014, Chomko 1975, Desmond 2018, Gates St-Pierre 2007, Griffits and Bonsall 2001, Langley et al. 2016, LeMoine 1994, Medina et al. 2018, Olsen 1979, Soffer 2004, Stone 2009, etc.). Though pioneered in the early/mid-twentieth century (e.g., Semenov 1964, Tyzzer 1936), experimental studies have only recently come to dominate how bone tools are interpreted, through the creation and use of replica bone tools for different tasks. Many early experiments were focused on functions of bone tools as weapons, and there is a wealth of contemporary literature surrounding weaponry and projectile uses of bone points (Arndt and Newcomer 1986, Bergman 1987, Bradfield and Brand 2013, Bradfield et al. 2020, Iovita & Sano 2016, Langley et al. 2020, Pokines 1998, Pétillon 2005, 2006, 2016, Rozoy 1992, Stodiek 1993, etc.). Here, we focus on non-weaponry functions, particularly everyday crafting practices used in a domestic context.
Determining bone tool functionality is possible through the study of microscopic abrasive changes which appear on the surface of bone tools during use. The basis for these studies is the concept of tribology. When a tool is used repeatedly in the performance of a specific task, abrasive wear patterns are produced in the form of changes to a bone tool’s surface topography (Campana 1989, LeMoine 1994). Because bone is softer than stone, it will retain traces of repeated contact with other materials as deformations to the tool’s surface at the point of contact: “[t]he plasticity and softness of bone allow rapid and characteristic recording of a given action. The original volume of the active end of the tool is more or less quickly deformed depending on the way the tool is used, the nature of the material being worked, its condition, as well as the duration of use”(Legrand and Radi 2008, 306). For example, if a bone tool is drawn repeatedly across a hide with granular inclusions, both the animal’s skin and any particulates on it will leave identifiable traces to the surface of the bone tool (Campana 1989, 63). These will differ from, say, traces developed through using a tool repeatedly to wedge wood (Campana 1989, 60–61). When examined at magnification, these traces can be distinguished from one another, even though the overall shape of the tool used in both of these processes may look nearly identical (e.g., a “point”). Archaeologists conducting experiments with replica bone tools can then compare the micro-topographic patterns produced on the experimental tools with patterns found on archaeological tools.
Experimentation has successfully documented use-traces associated with a variety of specific activities, including hide and skin piercing (Akhmetgaleeva 2017, Bradfield and Brand 2013, Campana 1989, Campana and Crabtree 2018, Griffits 2001, Legrand and Radi 2008, Wojtczak and Kerdy 2018,), scaling fish (Arrighi et al. 2016, LeMoine 1994), beading fish-heads and net making (Arrighi et al. 2016), use as hoes (Bradfield and Antonites 2018), in pottery smoothing (Buc 2011, Gates St. Pierre 2007, Struckmeyer 2011), termite-mound digging (Backwell and D’Errico 2001, Gardner 2019), corn-husking (Gates St-Pierre 2007), tattooing (Gates St-Pierre 2018), detaching limpets from rocks and gouging limpets out of their shells (Griffits and Bonsall 2001), as nose-bones (Langley et al. 2016), in basketry manufacture (Campana 1989, Campana and Crabtree 2018, Legrand and Radi 2008, Olsen 1979, Struckmeyer 2011), as clothing fasteners and bodkins, as implement handles, in splitting and shaving wood (Campana 1989), as hair-scrapers (LeMoine 1994), as musical instruments (Olsen 1979), in smoothing textiles while on a loom and gloss-linen polishing (Struckmeyer 2011), in sewing reeds (Arrighi et al. 2016), as billets (Olsen 1979), as hide compressors (Campana 1989, Martisius et al. 2020) and as wood wedges (Van Gijn 2005).
In other studies, use-trace analysis has focused on discerning changes associated with different worked material. Here the focus is on what the tool was used with, rather than what the tool was used for. These include general wet and dry material working (LeMoine 1994), tanned-skin working (Arrighi et al 2016, Campana 1989, Watson and Gleason 2016), dry-skin working (Buc 2011, Van Gijn 2005) fresh skin working (Buc 2011, Legrand and Radi 2008, Watson and Gleason 2016) bending birch bark (Arrighi et al 2016), in silica-rich-plant working (Buc 2011, Griffits 2001), rush-working (Buc 2011), smoked-hide piercing, bark piercing (Gates St-Pierre 2007, Watson and Gleason 2016), non-woody plant working, plant stripping/splitting (Griffits 2001), plant-scraping (Medina et al. 2018), sweat-scraping (Olsen 1979), plant fiber manipulation (Campana and Crabtree 2018, Soffer 2004, Stone 2009), nettle-working, sinew working (Stone 2009), bark peeling (Struckmeyer 2011, Van Gijn 2005), reed-working, hemp-working, digging (Van Gijn 2015), and yucca working (Watson and Gleason 2016).
Finally, some experimental studies have diagnosed use-wear associated with a particular motion of use or direction of force. These include longitudinal impact/pressure (Backwell et al. 2018, Bradfield 2012, Buc 2011), hafting, smoothing (Buc 2011), piercing (Campana 1989, Buc 2011, Christidou and Legrand 2005), fleshing/softening, punching/puncturing (Christidou and Legrand 2005), chiseling (Griffits 2001), indirect percussion, and rotation (Campana 1989, Legrand and Radi 2008).
The interpretation of archaeological bone tool microtopographies has successfully revealed a number of wide-scale, yet formerly imperceptible practices and processes. These include evidence for termite digging associated with Paranthropus robustus nearly two million years ago (Backwell and D’Errico 2001, Gardner 2019), seasonality in Palaeolithic economies, such as seasonal labor practices evident in on-site hide processing (Akhmetgaleeva 2017), targeted exploitation of wetland areas for plant-based crafting material (Arrighi et al. 2016), short-term occupation of special purpose camps devoted to shellfish processing (Griffits and Bonsall 2001), evidence for an advanced “maritime toolkit” – including fish gorges – among early California Palaeocoastal groups (Rick et al. 2001), the presence of textile weaving with a loom in Upper Palaeolithic Europe (Soffer 2004), a marked fluorescence/expansion of bone technologies in the Late Upper Palaeolithic signaling diversification in craft production and changing ideologies, economy, and role-specialization (Stone 2009), a high density of plant-working and clothing/bedding manufacture activities, indicating the presence of women on-site at a Neolithic camp (Van Gijn 2005), tanning processes using oak bark peelings (Struckmeyer 2011), the use of bones as garment pins/fasteners in the Chalcolithic (Campana and Crabtree 2018), very early perishable crafting, resin collecting, and/or personal ornamentation in Australia (Langley et al. 2016, Langley et al. 2021), and early musical instruments (Bradfield 2020) to name a few.
It is clear that a scientifically coherent approach to the study of bone tools – based on tribological principles and an examination of use trace patterning – can reveal a great deal about human activities in deep time, and in much more detail than lithic analysis alone. These shed light on formerly invisible actions, actors, and artefacts. As promising as these methods are, there remain significant challenges in the experimental design, analysis, and publication of use-wear results that have impeded research on bone tools.
One of the major difficulties facing researchers in this field is the ability to generate results of experimental studies in a reproducible way. To begin with, a standardized methodological protocol in bone tool analysis does not yet exist, and therefore analytical methods differ greatly between studies. These include scanning electron microscopy (SEM), light-microscopy, digital microscopy (e.g. Dino-Lite), metallurgical microscopes, hand-held lenses (e.g., a magnifying glass), macro photography, visual examination, etc. Methods can vary from magnifications of 10000× in the case of SEM to 10–250× with use of a light microscope, rendering comparisons of the resultant data untenable between studies. Even among studies using similar magnifications, publications often present the magnification used in different ways (i.e., the use of scale bars vs. listing the magnification used), and differences in depth of focus relative to each method may further distort images. Both the physical examination of tools and any published illustrations of use-traces are performed and recorded under different lighting conditions (e.g., in diffuse natural light vs. under a focused spotlight), as well as different lighting directions (e.g., above versus oblique), colors (e.g., yellow vs. white), and brightnesses (i.e., lumens). Differences in these conditions can vastly impact the way changes to bone tool surface topography appear. Published images are, of course, supplemented by written descriptions of use-traces, but here too, researchers may use different qualitative descriptions of use-wear. Some authors provide an index or glossary specifying terms used and their specific meanings, some broadly define descriptive terminology, and some do not.
Finally, sharing use-wear research involves the publication of two-dimensional images of bone tools. This can include illustrations, still captures from digital microscopes, SEM image files, photographs, etc. All of these are subject to further disparities in publication format, with factors such as fidelity, resolution/graininess, and the use of black and white, greyscale, or color, which vary by publication. Any published images of tools (via whatever method) are also inherently static, capable of showing a single aspect of tools’ surface topography in a single image. The use of different methods, lighting conditions, magnifications, resolutions, descriptive terms, and image fidelity in publications can further simplify or obscure the actual dynamic wear to surfaces which can occur on the surface of a three-dimensional object such as a bone tool.
This can be further compounded when archaeological bone tools are compared to vastly different materials and documentary media within a single study. Comparative datasets may include experimentally produced and used tools, taphonomic reference collections (physical or published), ethnographic data (varying in substance from written descriptions to photographs to access to physical reference collections), and the published results of other bone tool analyses. This latter category alone can include studies that foreground use-wear patterning, as well as studies from a time when use-trace analyses were far less developed, and when morphological identifications were still the “gold standard” in assessing bone tool function. Comparing use-wear on experimentally produced tools, which are available to the researcher in real time, to images or descriptions of tools in published studies can lead both to false-positive matches, or to erroneously discarding use-wear congruences due to obscure or different descriptive terms, perceived dissimilarities based on magnification, lighting angle, etc. Clearly, the use of all these different comparative datasets and methods together can result in a scientifically incoherent analysis, through the unwitting use of dissimilar, vague, or incommensurate comparative data.
The themes discussed here are broadly pervasive and, taken together, such methods are far from the tightly controlled parameters necessary to accurately determine past bone tool function. We therefore suggest a systematic protocol in assessing bone tool microtopography. First, we need to choose a scale of analysis. Magnification between 10–50× has broadly been established as a methodological standard for recording use-traces (Akhmetgaleeva 2017, Bradfield and Brand 2013, Bradfield 2014, Legrand and Radi 2008, Legrand and Sidéra 2007, Soffer 2004). At these magnifications whole-tool features are visible, there is less potential for image flattening or distortion which occurs at higher magnifications, color, polish (reflectiveness) and shape can be more accurately conveyed, and microscopes which work at these magnifications are widely available. We believe that most use-traces are best visualized at an observational and relational scale best suited to understanding the relationships between “zones of wear” (e.g., a polished base indicating a hand-held use, and an asymmetrically rounded tip indicating resharpening [Campana 1989]).
The need to examine bone tool topography at 10–50×, in color, and under controlled magnification and lighting parameters, using portable, easy-to-find equipment which is not prohibitively expensive, and then to be able to share results with other researchers in an interactive yet easily transmissible way allowed us to converge on a single technology: Reflectance Transformation Imaging, or RTI.
RTI was first developed at Hewlett Packard Laboratories in 2001 by Tom Malzbender and Dan Gelb. This is an imaging method which “…requires only images to generate high quality photorealistic renderings of a textured surface” (Malzbender et al. 2001). The basic premise of RTI involves lighting an object from a number of different directions, while a photo is taken at a fixed position above the object. Open-source software is then used to calculate “surface normals,” or vectors perpendicular to the object’s surface, to interpolate how light is reflected off the object’s surface at different lighting positions. These images are then synthesized into an output known as an RTI file. The resultant RTI file allows the viewer to interactively move a light source around a mathematical 3D map of the object, allowing even minute surface topography to be viewed in relief from a number of different positions (Figure 1).
All of this is done at very high resolution, as the surface detail is preserved on a per-pixel basis (Malzbender et al. 2001, Kinsman 2016a, Falcetano 2017a, b, c, Duffy et al. 2018). The RTI file can be used to further enhance surface topography for use-trace interpretation using specialized viewing applications, such as RTIViewer (Cultural Heritage Imaging 2013 [available at CulturalHeritageImaging.org]). The RTI viewer allows lighting positions to be further manipulated by applying a variety of rendering modes, and creating visualizations which can enhance surface inspection of objects. For example, rendering modes such as multi-light enhancement produce extreme raking-oblique light on the tools’ surface, without the loss of shadow detail (which is problematic in conventional object photography lighting). Additionally, applying a surface normals rendering mode allows for further critical examination of a surface at a pixel level.
RTI has a number of benefits over other imaging techniques. To begin with, RTI offers advantages over conventional photography. While using advances in camera technology, RTI allows for “computational photography” and surmounts the limitations of photographs as static images of an object (Fornaro and Chiquet 2019). With conventional archaeological object photography, the photographer chooses the spectral quality of lighting and edits the lighting position before camera capture to enhance diagnostic features seen on a subject from the camera position. Here, the photographer chooses to show (or is restricted to reveal) details given by lighting opportunities. The final publication photograph is then presented as objective, though it is arguably subjective. With RTI photography the relationship between the human eye and camera position is infinite and surpasses the limitations of the human eye, camera, and lighting of conventional photographs. It captures a multitude of views and can be critically understood to be a scientific image. When compared with photogrammetry, RTI works much better on reflective, smooth, or polished surfaces, which tend to make point-tracking in photogrammetry very difficult. This is crucial as many bone tools will exhibit use-wear in the form of smoothed areas of high-gloss or polish. When compared with GIGAmacro (a high-resolution digital image bitmap), RTI capture times are much quicker, and retain more information per-pixel as a result of multiple lighting angles (Falcetano 2017a, b, c). RTI also offers a much higher resolution and better visual representation of an object’s surface than does 3D scanning, and is far more portable, less expensive, and works on a scale better for most interpretive analyses than does an SEM. Further, RTI captures are non-contact, and do not require covering the object in gold as with an SEM. When compared with traditional microscopy, RTI can capture dynamic images of objects as opposed to viewing them in isolation from a single perspective (Newman 2015). Of course we do not suggest that photos, microscope stills, written descriptions, etc. be eliminated from use-wear studies; rather we suggest that they be used in concert with RTI, in order to show and describe use-wear as fully as possible and to best advantage.
RTI has a proven record in successfully tracking subtle changes in objects’ surface topography. For example, it has been used to find minute damage to the surface of paintings (Manfredi et al. 2014), to isolate and read ink on carbonized papyri (Piquette 2017), and for viewing original designs on degraded metal artefacts (Roberts Thompson and Williams 2016). RTI has already proven useful in viewing modified bone surfaces, such as identifying butchery marks on bone and antler artefacts (Mytum and Peterson 2018). It has also been used in legal/forensic contexts, to analyze and distinguish specific trauma and tool marks occurring on the surface of bones (Clarke and Christensen 2016). Images produced by RTI clearly meet a high/evidentiary standard of fidelity to the bone analyzed, as the authors suggest it for legal use in laboratory, courtroom, and forensic environments (Clarke and Christensen 2016).
There are two main methods for RTI capture: H-RTI and dome RTI. The most common of these is the highlight method (H-RTI), wherein a hand-held light source (such as a flashlight) is moved into different hemispherical positions around an object in a darkened room, and a photograph is taken at each lighting position (Mytum and Peterson 2018). A fixed distance between the object and flashlight can be determined by use of a string cut to a specific length (Mytum and Peterson 2018). Small reflective ball bearings are placed in-frame with the object, and the position of lighting for each capture is recorded by calculating the position of the light reflection on the three-dimensional sphere. Software is then used to calculate the position of the light relative to the object, allowing for a seemingly three-dimensional topographic rendering of the object’s surface (Duffy et al. 2018, Macdonald 2017, Kinsman 2016a).
In contrast to H-RTI, dome RTI captures are performed with use of a prefabricated dome structure which is placed over the object. This is generally done in museums and other settings where the portability of an H-RTI set-up is not important. This dome contains a number of lights affixed at different locations over the object in a hemispheric position, and are automated to turn on individually with each camera capture (Duffy et al. 2018). This set-up obviates the need for a darkened room, a flashlight, and ball bearings, as the dome itself blocks any other light source. As the positions of the lights are pre-determined, the images can be automatically synthesized into an RTI file.
Each of these RTI methods has benefits and drawbacks. For example, the elements of an H-RTI kit are inherently portable, consisting primarily of camera equipment and a light. This means H-RTI can be used to document artifacts on site; it is particularly well suited for capturing objects in situ (providing the lighting conditions can be met), and has even been used underwater (Selmo et al. 2017). This is crucial in situations where artefacts or collections are delicate, in remote locations, in precarious geopolitical situations, or non-exportable, and creates a viable working model when the object being recorded is physically absent (Hunziker-Rodewald and Fornaro 2019). However, H-RTI can be time consuming, and if the object moves during capture the process has to be started all over again. To date, the single published application of RTI to archaeological bone tools used a H-RTI capture method consisting of 36 photographs taken with a flashlight at different positions around the object, using a standard camera set-up. Though the dynamic nature of RTI files was found to be very useful, here the utility of H-RTI was limited to pre-selecting artefacts which would merit further analysis via other methods, time-consuming when compared to other methods, and only able to capture details already visible to the naked eye (Newman 2015).
Dome RTI capture is far less tedious and time consuming with its automated lighting and image processing. Dome RTI controls for many of the variables which have hampered comparability in other studies, as there is an automatic control over the lighting distance, color, direction, and intensity (brightness). This also eliminates the need for a string in checking the distance between the object and the hand-held light source, making captures far less tedious, and reducing inadvertent bumping or moving of the object, which renders the capture sequence unusable. The object is protected by the dome during image capture, and post-processing can be easily automated without the need for reflective ball-bearings. This, however, does not allow for editing photos to maximum advantage prior to processing. Additionally, the non-portability of most RTI domes limit its use outside of static museum and/or conservation settings. In view of the limitations of RTI presented by each method of capture, we chose to tailor our own system to use the benefits of H-RTI and dome RTI, at the scale best suited for rendering bone tool use-wear.
Foremost, we realized that a standard camera lens would not capture the level of bone tool surface detail necessary for comparative analysis, but replacing a standard camera lens with a macro lens would allow us to capture images at the necessary magnification and resolution. Due to the smaller sensor used and larger image captured with the macro lens, the image is easily scalable. Macro lenses also provide a fixed focal length, which is perfect for use in a lighting dome and can be used in a number of other imaging techniques (photogrammetry for example). Macro lenses are also useful in conventional photography as they can focus to infinity. When using a macro lens, focal length influences how close one needs to be to the subject: this is known as working distance. With a greater focal length, it is possible to be farther from the subject and still achieve the effect of extreme closeness. Smaller focal lengths allow a very close working distance while easy focus acquisition is maintained. This has benefits over the use of a microscope, with which a very small area is covered during image capture. Areas outside the central view are often out of focus, unless computational focus depth stacking is utilized. For this project we chose a Nikon 60 mm Micro Nikkor lens with manual focus giving a 1:1 reproduction, and a high MTF performance (Modulation Transfer Function or “MTF” is a measurement of the optical performance potential of a lens). This allows for minimum distortion and accuracy of color value recording of fine details on the camera sensor.
Since documenting bone tool microtopography happens both on-site and in museum collections, we wished to develop a technology which was both robust and portable enough to use in all settings, while also allowing for the best possible capture outcomes. This seemed most achievable through use of a portable lighting dome (Figure 1). Foremost, this would eliminate the time-consuming nature of the H-RTI capture, while simultaneously allowing for more lighting positions (50+) to construct a more detailed final RTI file. We combined the systematic and thorough lighting conditions provided by a dome (in addition to the object being completely protected and eliminating ambient light) with the benefits of H-RTI, which include portability, cheapness, etc. In field settings it is not always possible to find circumstances akin to the “darkened room” necessary for standard H-RTI. Furthermore, we needed a dome which did not require an electrical outlet as its power source, and which was not likely to be damaged by excavation conditions such as dust or debris (at least not any more than standard camera equipment). However, we did wish to retain the use of ball bearings as an aspect H-RTI, allowing us more control in editing photos before they are compiled into an RTI file. Capturing these photos in RAW crucially allowed us to lighten/darken images to best advantage before running them through the RTI builder (available at CulturalHeritageImaging.org), an advantage over most domes’ standard automatic processing. Since it is the RAW data which is captured by the camera sensor, it is possible to repeatedly change and reformat; JPEG captures, for example, don’t allow for this degree of choice and control. This is absolutely necessary for high fidelity reproduction of true color, and also editing exposure, shadows, highlights etc.
Though there are a number of open source RTI designs available (Falcetano 2017a, b, c, Pawlowicz 2016), our design (Figure 1) was based on Kinsman 2016b, with a number of modifications (please see https://github.com/BeebBenjamin/RTIPy for specifications). We used all the same size LED bulbs in a white light for best color fidelity, and in terms of software to run the dome capture sequence, we developed a lightweight Python-based terminal application to send different switch commands to an arduino (Figure 1a), which allowed us to control our own lighting settings from a laptop. Because we use the USB/Serial cable, the entire dome can be powered from the laptop, with the added possibility of using a 9V battery in future. Instead of a reed switch we used an infrared LED with a timed pulse to trigger the camera; this in theory could be programmed to include other camera pulses, but at the moment it only works with Nikons. We also added a 30 second pause before beginning the capture sequence, to allow any vibrations from the camera set-up to abate (Figure 1b).
The flexibility of our design showcases the modularity of RTI as a method. Because the LED’s are not soldered into the dome, we can run any number of lighting positions (e.g. 30 rather than 72) for a quicker capture. We can also swap out our white LEDs for ones which emit different wavelengths of light. Because of this we have an in-built capacity to make our RTI capture multi-spectral by changing the LEDs and using a modified camera/lens. Further, because we use an H-RTI technique within a lighting dome, it is theoretically possible to 3D print multiple setups and try new things very quickly by altering our virtual dome design. Should a researcher have questions relating to smaller magnification-specific microtopographies, a dome could be printed to fit a small light microscope (such as a Dino-Lite) rather than a camera. The 3D printed dome is lightweight and portable, and can be placed against non-level surfaces by using a magic arm on the camera. Adding a lens clip would allow one to actually attach the dome to the end of the macro lens, and different domes could be printed to accommodate different lens widths. This is particularly useful in field settings where artefacts and features do not always occur at level planes. Particular research questions can certainly be investigated by incorporating other techniques and dome prints, such as the use of higher magnifications with Micro RTI (Reilly et al. 2021).
Tools were first examined visually, photographed using a Nikon D7100 camera and a Nikon 60 mm Micro Nikkor lens, then examined and photographed with a Dino-Lite digital microscope (AM3113T) before running RTI captures. We use a modified version of Duffy et al. (2018) for our basic RTI protocols (supplementary materials).
Rather than choosing archeological or experimental tools, we have chosen a selection of ethnographic bone tools with known uses and articulations. These represent tools used in a variety of “real world” scenarios, and will have developed microtopographic use-wear throughout their life history. A selection of tools was chosen from the Phoebe A. Hearst Museum of Anthropology (PAHMA) collections at UC Berkeley, and are representative of different non-weapon perishable crafting use-trace pattern development. We chose tools with a variety of uses in order to test whether RTI could effectively record microtopographic patterning related to the tools’ functions.
In order to illustrate bone tool surface detail captured by RTI, we have included conventional photographs, digital microscope images, and stills from RTI files which show different features at a number of magnifications. We also discuss filters which can be applied to RTI files, and examples of their specific utility in interpreting bone tool use-wear. The original RTI files can be downloaded from https://doi.org/10.7910/DVN/OLZ7AX.
First, we wish to show the dynamic nature of the RTI file, and how it can be manipulated to reveal different features through subtle changes in lighting position (Figure 2). In order to show how the virtual light source is manipulated, we show four views of an antler tool used in woodwork-finishing; specifically, in polishing projecting edges and angles of carved wooden items. They illustrate how changes in the direction of lighting illuminate different surface features, including the presence of rough/smooth areas, longitudinal and diagonal use-trace striations, concretions, etc. (Figure 2a–d). Were any one of these photos to represent the tool in a publication, the reader would have a different (and incomplete) representation of how the tool’s surface actually looks.
Next, we can see how RTI captures compare with standard macro photos and digital microscope stills. Figure 3 shows a bone tool made from a deer metatarsal which was used in making coiled baskets (Native name: “tculla”). Image Figure 3a is a standard macro photo with a scale bar, and Figure 3b–c show digital microscope captures at 50× and 230×. Figure 3d–e show the same object’s RTI capture viewed with a 15% zoom, Figure 3e–f with a 28% zoom, and Figure 3h–i with a 41% zoom. Because of the greater degree of control in image processing provided through our design, we were able to enhance the uncompiled images’ exposure and contrast before running them through the RTI builder (Figure 3d–g). This allowed us to make fine surface topographies more readily visible by enhancing the contrast between areas of different elevation on the tools’ surface by modifying white balance, clarity, and exposure settings. This enables a clear visual representation of use traces not possible with standard microscopy and still photos (Figure 3a–c). This also shows the benefit of using RTI in conjunction with other methods. While Figure 3a–c are useful in showing static images of surface topography, RTI (Figure 3d–i) allows for the examination of these use-traces in seeming three dimensions from a number of different lighting positions, in relationship to one another, at a wide variety of magnifications (zooms), across a broad portion of the tool’s surface in a single image file. Interesting use-wear captured by a still photo or microscope can be more fully examined with RTI, viewed in relation to other use-wear and tool features, and shared with other researchers without a loss of fidelity.
We also show how RTI can document bone tools which exhibit a high degree of polish developed through use. As previously mentioned, highly reflective areas can be difficult to document using point-tracking in photogrammetry. Polish is also difficult to capture using flash photography and light microscopy, as the light source which illuminates the object can reflect back into the lens, causing flares and extremely bright “blown out” images. The object considered here is a skin scraper: repeated contact with skins and hides has developed a high polish to the working edge of the tool. Figure 4a shows a conventional photograph; note how the polish to the tool’s lower right edge is not evident, as the object was photographed in darker conditions to exclude flare from this highly polished surface. Figure 4b shows a Dino-Lite still of the object’s blade taken at 50× magnification. In this image the tool appears much lighter in color, and in places the image has been washed out due to glare from the light source on the reflective surface. A still from the RTI file’s standard view at 40% zoom (Figure 4c) show how RTI is able to capture the object’s color with greater fidelity, and to show areas of high gloss/polish, smoothness, and faint striations clearly and without glare or color distortion. Here we show how the “diffuse grain” filter can be applied to bone tools: Figure 4d shows the same portion of the tool (at 40× zoom), with an increased representation of height and depth on the tool’s surface. This tool allows researchers to view topography in the form of peaks/depressions, and “enhances the perception of surface shape features of the subject for interpretive purposes” (Cultural Heritage Imaging 2013). This is also a valuable tool in use-wear interpretation: when the part of the tool being analyzed is similar in color but has differences in surface topography (e.g. Figure 4c), this rendering mode can be applied to better understand the topography or surface features and use-wear, where a polished area (indicating a surface of repeated contact) ends and a non-worked surface begins (e.g. Figure 4d).
Next we show a tool which possesses multiple attributes that are difficult to document; specifically, high-polish and small, extremely fine use-traces present on the same part of the tool. This tool is a bone awl made from a deer ulna and used for making basketry. Similar to the skin-scraper, the photo of the basketry tool does not effectively capture areas of high polish (Figure 6a), and the reflectivity of the polish makes accurately capturing microscopic striations difficult at 20× magnification (Figure 6b). The RTI file, however, shows both the smoothness of the polished area and the microscopic diagonal striations clustered atop it along the axis of the tool, first at a 15% zoom (Figure 6c) and then at a 36% zoom (Figure 6d).
Finally, we show how use of the “specular enhancement” filter allows for the viewing of surface shape/topography, and enhances the reflectiveness of smooth surfaces. Figure 5a illustrates how the default view in the RTI viewer shows a true-color dynamic rendering of the surface topography, including decorative incisions, concretions, etc. Figure 5b shows how using the “specular enhancement” filter provides an enhanced visualization of the lower portion of the tool’s blade. This is useful for identifying/discerning areas of high polish/smoothing, and contrasting them with adjacent areas with more varied surface topography. Again, this a crucial step in identifying use-trace patterning, as areas of higher polish occur on portions of the tool which repeatedly came into articulation with the hand or worked materials. By using the specular enhancement filter, researchers are easily able to identify wear-foci in ways not possible with standard image capture methods.
RTI clearly offers advantages in representing minute surface topographies on bone tools used in perishable crafting. While specific use-wear analyses are beyond the scope of the present study, it is possible not only to see minute surface topography in three dimensions from multiple angles, but also to track the ordinal nature of changes (e.g., polish overlain by scratching). This is important in distinguishing tool formation strategies (i.e., initial shaping/resharpening) from use-wear, and these from any post-depositional damage. Additionally, researchers can apply a number of filters within the RTI viewer to provide an enhanced view of topographical features. “Diffuse grain” visually enhances changes in height and depth, while “specular enhancement” enhances surface topography and reflectiveness, and are both valuable for identifying use-wear foci. Because RTI uses a camera lens it can capture a larger area than microscope stills, which makes it possible to see the relationships between “zones of wear” on a tool (e.g., polish to the base indicating hand-held use and tip striations indicating direction of use). A greater degree of flexibility in image post-processing allows researchers to enhance the visibility of use-wear by modifying images’ sharpness, exposure, white balance, contrast, etc., before they are compiled into an RTI file. Because of the modularity of the RTI design we present here, it is further possible for researchers to choose their preferred scale of analysis (e.g., macro photography, use of a light microscope in image capture, etc.) in order to address specific research questions or objectives. RTI can also be combined with other methods in order to yield the most thorough record of a surface: in particular, newly published applications of RTI with photogrammetry are very promising (Solem and Nau 2020). Perhaps most importantly, researchers can compare diagnostic use-traces between bone tools in “real-time” without having to have the objects physically present. This allows researchers to interpret bone tool use-wear in a scientific way, with a greater degree of control over variables in capture than perhaps any other currently available method of record, and to share these captures without loss of fidelity.
RTI provides a virtual working model of microtopographic surface features related to bone tool use. RTI captures provide a dynamic, scalable view of minute use-wear not possible with standard photography or microscopy. It is a technology which is open-source, easy to modify, inexpensive, portable, and non-invasive. Because of its control over variables in capture, ability to capture use-wear not easily recorded by other techniques, and the dynamism of the images produced, RTI can be used to document minute use-trace patterning on bone tools under tightly controlled parameters. Since well-documented ethnographic collections and experimental bone tools are not always physically accessible, RTI files of these tools shared via a web platform can serve as a remotely accessible dataset. These can then be compared with archaeological bone tools to determine best-fit use scenarios. In this way, RTI enables researchers to move toward bone tools as a proxy for otherwise invisible crafting practices and organic products (such as weaving, basket making, netting, etc.), through identifying diagnostic use-wear patterning on archaeological bone tools.
Understanding the role of perishable crafting can tell us a great deal about a group’s technological aptitudes, and provides a much more comprehensive framework for understanding the breadth of people’s activities, roles, and customs in prehistoric archaeological contexts. As the archaeological record is taphonomically biased toward stone tools and hunting-related activities, understanding bone tool use-wear is crucial in developing a more accurate picture of how the remote past was actually lived, and by whom.
This protocol is a modified version of that suggested by Duffy et al. (2018), and uses the RTI builder and RTI viewer available at culturalheritageimaging.org. For each object we took standard photos, macro photos, and Dino-Lite microscope captures. What follows is a description of the RTI capture sequence, start to finish.
We began by placing a square of black velvet on the working surface (table/desk), and then placed the tool atop the black velvet. This could also be done on a non-level surface (such as a bone tool in an in-situ archaeological context), but would require editing the background in post-processing. Reflective ball bearings were adjusted to the height of the tool using clay (blue tac). It was important to make sure that no clay stuck to the top of the ball bearings visible to the camera, as this could interfere with the capture. Furthermore we learned that only highly reflective and spherical ball bearings can be used; some we ordered were of poor quality and dented/not shiny, so this also interfered with some initial captures. Once the object and ball bearings were in place, the camera was mounted and adjusted above the object to bring it into focus (Nikon 60 mm Micro Nikkor lens). It was useful to have a glass cleaning cloth to hand to wipe the top of the ball bearings once in place.
Camera settings depended on the color of the tool being photographed; for all tools, we set the aperture to F16. For light and medium-colored tools (qualitatively assessed) we used a shutter speed of 1/40th and an ISO of 500. If the photo looked too dark, we adjusted the shutter speed to 1/30th or 1/20th. For darker colored tools we used a shutter speed of 1/15th and an ISO of 640.
For the first photo (not compiled with the RTI captures), a scale bar was placed in frame, also using clay to bring it to the height of the object for accuracy in measurement. This allowed us to accurately scale any RTI files. Because the ball bearings are of a known size these can also function as an in-built scale. Next, we ensured that all LEDs were pressed firmly into the holes drilled in the RTI dome. Since we wished to use all of the 72 automatic lighting positions, the placement of the specific bulbs did not matter. However, if one wished to run, say, only 30 out of the 72 lighting positions, it would be necessary to move the first 30 LEDs to the desired positions within the dome, and to then set the RTI program to run to 30 shots. After all LED bulbs were secured in the dome, the dome was placed above the object, with the camera directly above the opening. We then turned on the infrared remote for the camera, which allowed us two minutes to begin the capture sequence. To avoid any ambient light, a large piece of black velvet was placed over the entire set-up, covering both the camera and the dome to ensure total darkness in capture. Once the camera was in position over the object and the velvet had been placed, we opened the RTI program on the laptop (MedusaPy.py). Here there is an additional check to make sure the object is visible and in focus, lighting the object from one LED position for one minute. This also allows time for any vibrations from setting up the camera to abate and ensure stillness in capture. RTI capture can then be started, and the dome will automatically light to 72 positions with a photograph taken at each position. After running the capture sequence, we imported the RAW files from the camera’s SD card into a folder (with the corresponding object and capture number) on the computer desktop.
Because we captured in RAW the resulting folders were very large. After each capture we moved the folder containing the images from the desktop onto an external hard drive to save space. To “assembly line” the process and enable more RTI captures we ran one capture after another, and then waited to compile them until afterward. We found it was a good idea to run one or two captures at random through the RTI builder every evening to make sure everything was working as planned. By doing this it is possible to catch, for example, if there is something stuck to or obscuring the ball bearing, maybe a capture is out of focus, and/or any other issues in case a capture has to be redone.
When we were ready to compile an RTI capture, we opened all the photos of a given capture sequence (minus the initial photo with the scale bar) in Adobe Photoshop, and deleted any which were too dark. For the remainder, we chose the most appropriate tools for showing the surface relief to best advantage on a photo by photo basis: changing the white balance, increasing exposure and clarity where necessary, etc. Then, we exported all the photos as .jpegs with maximum quality into a subfolder (jpeg-exports). Next we open the RTI builder program, and for “project name” we select the folder with the exported jpegs. Next, under “operation sequence” we chose the highlight based PTM filter, choose “open folder”, and choose the appropriate desktop folder.
All the images comprising the final file will then open in the RTI builder program. At this stage, one can also tick boxes to get rid of any additional dark/bad images. Next, we chose an image with bright white reflection in at least one of the ball bearings and visible sphere peripheries, using the red/green squares to outline, as exactly as possible, the ball bearing. Now the program will show a bluish topographic map of the sphere: one must drag and move a circle to surround (as accurately as possible) the outside of the ball bearing, with the cross in the center of the circle. Next, after choosing whether to crop the image, we chose “highlight-detection”, and ran the program.
The authors would like to thank Nick Barton, Alison Roberts, The Ashmolean Museum, Abdeljalil Bouzouggar, Louise Humphrey, Hayley Mountford, Clara Rice, The Phoebe A. Hearst Museum of Anthropology, Natasha Johnson, the Inupiat, Yupiat, Pomo, and Me-wuk (Miwok) peoples, Julius Koll, and Jane Desmond.
The authors have no competing interests to declare.
Akhmetgaleeva, NB. 2017. Hide polish of worked bone tools from East European Late Upper Paleolithic sites of Byki (Russia). Quaternary International, 427: 225–229. DOI: https://doi.org/10.1016/j.quaint.2015.12.022
Antonites, AR, Bradfield, J and Forssman, T. 2016. Technological, functional and contextual aspects of the K2 and Mapungubwe worked bone industries. African Archaeological Review, 33(4): 437–463. DOI: https://doi.org/10.1007/s10437-016-9233-z
Arndt, S and Newcomer, M. 1986. Breakage Patterns on Prehistoric Bone Points: An Experimental Study in Studies in the Upper Palaeolithic of Britain and Northwest Europe. BAR International Series, 296: 165–173.
Arrighi, S, Bazzanella, M, Boschin, F and Wierer, U. 2016. How to make and use a bone “spatula”. An experimental program based on the Mesolithic osseous assemblage of Galgenbühel/Dos de la Forca (Salurn/Salorno, BZ, Italy). Quaternary International, 423: 143–165. DOI: https://doi.org/10.1016/j.quaint.2015.11.114
Backwell, L, Bradfield, J, Carlson, KJ, Jashashvili, T, Wadley, L and D’Errico, F. 2018. The antiquity of bow-and-arrow technology: Evidence from Middle Stone Age layers at Sibudu cave. Antiquity, 92(362): 289–303. DOI: https://doi.org/10.15184/aqy.2018.11
Backwell, LR and D’Errico, F. 2001. Evidence of termite foraging by Swartkrans early hominids. Proceedings of the National Academy of Sciences, 98(4): 1358–1363. DOI: https://doi.org/10.1073/pnas.98.4.1358
Berecz, B, Cyrille, M, Casselbrant, U, Oleksak, S and Norholt, H. (2020). Carrying human infants – An evolutionary heritage. Infant Behavior and Development, 60: 101460. DOI: https://doi.org/10.1016/j.infbeh.2020.101460
Bergman, CA. 1987. Hafting and use of bone and antler points from Ksar Akil, Lebanon. In: La main et l’outil, Manches et emmanchements préhistoriques, actes du Colloque de Lyon, 1984, Stordeur, D (ed.), 117–126. Paris: CNRS Éditions.
Biddittu, I and Celletti, P. (2001, October). Plio-Pleistocene proboscidea and Lower Palaeolithic bone industry of southern Latium (Italy). In: The world of elephants. Proceedings of the First International Congress. Rome: Consiglio Nazionale delle Ricerche, 91–96.
Bradfield, J. 2014. Pointed bone tool technology in Southern Africa. Azania: Archaeological Research in Africa, 50(1): 148–149. DOI: https://doi.org/10.1080/0067270X.2014.993222
Bradfield, J. 2020. Use-Wear Analysis Brings “Vanished Technologies” to Light. African Archaeological Review, 37(4): 615–626. DOI: https://doi.org/10.1007/s10437-020-09412-8
Bradfield, J and Antonites, AR. 2018. Bone hoes from the Middle Iron Age, Limpopo province, South Africa. Quaternary International, 472: 126–134. DOI: https://doi.org/10.1016/j.quaint.2017.10.028
Bradfield, J and Brand, T. 2013. Results of utilitarian and accidental breakage experiments on bone points. Archaeological and Anthropological Sciences, 7(1): 27–38. DOI: https://doi.org/10.1007/s12520-013-0136-5
Buc, N. 2011. Experimental series and use-wear in bone tools. Journal of Archaeological Science, 38(3): 546–557. DOI: https://doi.org/10.1016/j.jas.2010.10.009
Campana, DV. 1989. Natufian and Protoneolithic Bone Tools: The Manufacture and Use of Bone Implements in the Zagros and the Levant. Douglas V. Campana. BAR International Series 494. Oxford: British Archaeological Reports, 1989. DOI: https://doi.org/10.30861/9780860546320
Campana, DV and Crabtree, PJ. 2018. Bone implements from Chalcolithic Tepecik-Çiftlik: Traces of manufacture and wear on two classes of bone objects recovered from the 2013 excavation season. Quaternary International, 472: 75–83. DOI: https://doi.org/10.1016/j.quaint.2017.11.042
Chomko, SA. 1975. Bone “Awls” and utilized antler tines from Arnold research cave, 23Cy64, Missouri. Plains Anthropologist, 20(67): 27–40. DOI: https://doi.org/10.1080/2052546.1975.11908724
Christidou, R and Legrand-Pineau, A. 2005. Hide working and bone tools: experimentation design and applications. In From Hooves to Horns, from Mollusc to Mammoth. Manufacture and Use of Bone Artefacts from Prehistoric Times to the Present: Proceedings of the 4th Meeting of the ICAZ Worked Bone Research Group, at Tallinn, 26–31 of August 2003, 15: 385–396.
Clarke, S and Christensen, AM. 2016. Reflectance transformation imaging (RTI) of saw marks on bones. Journal of Forensic Radiology and Imaging, 7: 33–37. DOI: https://doi.org/10.1016/j.jofri.2016.11.004
Croes, D. 1997. The North-Central cultural dichotomy on the Northwest Coast of North America: its evolution as suggested by wet-site basketry and wooden fish-hooks. Antiquity, 71(273): 594–615. DOI: https://doi.org/10.1017/S0003598X00085355
Cultural Heritage Imaging. 2013. Guide to RTIViewer. Cultural Heritage Imaging and Visual Computing Lab, ISTI – Italian National Research Council. http://culturalheritageimaging.org/What_We_Offer/Downloads/rtiviewer/RTIViewer_Guide_v1_1.pdf.
Desmond, A, Barton, N, Bouzouggar, A, Douka, K, Fernandez, P, Humphrey, L, Morales, J, Turner, E and Buckley, M. 2018. ZooMS identification of bone tools from the North African Later Stone Age. Journal of Archaeological Science, 98: 149–157. DOI: https://doi.org/10.1016/j.jas.2018.08.012
Duffy, SM, Goskar, T, Backhouse, P and Kennedy, H. 2018. Multi-light Imaging Highlight-Reflectance Transformation Imaging (H-RTI) for Cultural Heritage. Historic England. https://historicengland.org.uk/images-books/publications/multi-light-imaging-heritage-applications/.
Falcetano, K. 2017a. An introduction to automated RTI. Latis Labs. https://labs.dash.umn.edu/aisos/an-introduction-to-automated-rti/.
Falcetano, K. 2017b. Building an Automatic RTI Dome: Wrapping It Up. Latis Labs. https://labs.dash.umn.edu/aisos/building-an-automatic-rti-dome-wrapping-it-up/.
Falcetano, K. 2017c. Building an automated RTI dome. Latis Labs. https://labs.dash.umn.edu/aisos/building-an-automated-rti-dome/.
Gates St-Pierre, C. 2018. Needles and bodies: A microwear analysis of experimental bone tattooing instruments. Journal of Archaeological Science: Reports, 20: 881–887. DOI: https://doi.org/10.1016/j.jasrep.2017.10.027
Gates St-Pierre 2007 Pierre, CG. 2007. Bone Awls of the St. Louis Iroquoians: A Microwear Analysis. In Bones as tools: Current methods and interpretations in worked bone studies, 107–118. British Archaeological Reports. DOI: https://doi.org/10.30861/9781407300344
Griffits. 2001. International Council for Archaeozoology. Worked Bone Research Group. Meeting. Bone tools from Los Pozos. In: Crafting bone: Skeletal technologies through time and space: Proceedings of the 2nd meeting of the (ICAZ) worked bone research group, Budapest, 31 August – 5 September 1999. British Archaeological Reports.
Griffits, J and Bonsall, C. 2001. Experimental Determination of the Function of Antler and Bone Bevel-ended Tools from Prehistoric Shell Middens in Western Scotland. In: Crafting bone: Skeletal technologies through time and space: Proceedings of the 2nd meeting of the (ICAZ) worked bone research group, Budapest, 31 August – 5 September 1999, 207–220. British Archaeological Reports.
Hunziker-Rodewald, R and Fornaro, P. 2019. RTI images for documentation in archaeology: The case of the Iron Age female terracotta figurines from Buṣayra, Jordan. Journal of Eastern Mediterranean Archaeology & Heritage Studies, 7(2): 188. DOI: https://doi.org/10.5325/jeasmedarcherstu.7.2.0188
Hurcombe, L. 2008. Organics from inorganics: Using experimental archaeology as a research tool for studying perishable material culture. World Archaeology, 40(1): 83–115. DOI: https://doi.org/10.1080/00438240801889423
Iovita, R and Sano, K. 2016. Thirty Years of Experimental Research on the Breakage Patterns of Stone Age Osseous Points. Overview, Methodological Problems and Current Perspectives. In: Multidisciplinary approaches to the study of Stone Age weaponry, 47–63. Springer. DOI: https://doi.org/10.1007/978-94-017-7602-8
Kinsman, T. 2016a. An easy to build reflectance transformation imaging (RTI) system. Journal of Biocommunication, 40(1). DOI: https://doi.org/10.5210/jbc.v40i1.6625
Kinsman, T. 2016b. Shoot Super Detailed Macro Photographs With An RTI Camera Rig. [online] Make: DIY Projects and Ideas for Makers. Available at: <https://makezine.com/projects/shoot-super-detailed-macro-photographs-with-an-rti-camera-rig/> [Accessed 26 November 2020].
Langley, MC, Amano, N, Wedage, O, Deraniyagala, S, Pathmalal, M, Perera, N, Boivin, N, Petraglia, MD and Roberts, P. 2020. Bows and arrows and complex symbolic displays 48,000 years ago in the South Asian tropics. Science Advances, 6(24): eaba3831. DOI: https://doi.org/10.1126/sciadv.aba3831
Langley, MC, Balme, J and O’Connor, S. 2021. Bone artifacts from Riwi cave, south-central Kimberley: Reappraisal of the timing and role of osseous artifacts in northern Australia. International Journal of Osteoarchaeology. DOI: https://doi.org/10.1002/oa.2981
Langley, MC, O’Connor, S and Aplin, K. 2016. A >46,000-year-old kangaroo bone implement from carpenter’s gap 1 (Kimberley, northwest Australia). Quaternary Science Reviews, 154: 199–213. DOI: https://doi.org/10.1016/j.quascirev.2016.11.006
Legrand, A and Radi, G. 2008. Manufacture and Use of Bone Points from Early Neolithic Colle Santo Stefano, Abruzzo, Italy. Journal of Field Archaeology, 33(3): 305–320. DOI: https://doi.org/10.1179/009346908791071196
LeMoine, GM. 1994. Use wear on bone and antler tools from the Mackenzie delta, Northwest Territories. American Antiquity, 59(2): 316–334. DOI: https://doi.org/10.2307/281935
Macdonald, L. 2017. Reflectance transformation imaging. Digital Techniques for Documenting and Preserving Cultural Heritage. DOI: https://doi.org/10.5040/9781641899444.017
Malzbender, T, Gelb, D and Wolters, H. 2001. Polynomial texture maps. Proceedings of the 28th annual conference on Computer graphics and interactive techniques – SIGGRAPH ‘01. DOI: https://doi.org/10.1145/383259.383320
Manfredi, M, Bearman, G, Williamson, G, Kronkright, D, Doehne, E, Jacobs, M and Marengo, E. 2014. A new quantitative method for the non-invasive documentation of morphological damage in paintings using RTI surface normals. Sensors, 14(7): 12271–12284. DOI: https://doi.org/10.3390/s140712271
Martisius, NL, Welker, F, Dogandžić, T, Grote, MN, Rendu, W, Sinet-Mathiot, V, Wilcke, A, McPherron, SJ, Soressi, M and Steele, TE. 2020. Non-destructive zooms identification reveals strategic bone tool raw material selection by Neandertals. Scientific Reports, 10(1). DOI: https://doi.org/10.1038/s41598-020-64358-w
Medina, M, López, L and Buc, N. 2018. Bone tool and tuber processing: A multi-proxy approach at Boyo Paso 2, Argentina. Antiquity, 92(364): 1040–1055. DOI: https://doi.org/10.15184/aqy.2018.93
Mytum, H and Peterson, JR. 2018. The application of reflectance transformation imaging (RTI) in historical archaeology. Historical Archaeology, 52(2): 489–503. DOI: https://doi.org/10.1007/s41636-018-0107-x
Newman, SE. 2015. Applications of reflectance transformation imaging (RTI) to the study of bone surface modifications. Journal of Archaeological Science, 53: 536–549. DOI: https://doi.org/10.1016/j.jas.2014.11.019
Olsen, SL. 1979. A study of bone artifacts from grasshopper Pueblo, AZ P:14:1. KIVA, 44(4): 341–373. DOI: https://doi.org/10.1080/00231940.1979.11757924
Pawlowicz, L. 2016. Building the RTI system. Reflectance Transformation Imaging For Lithics|Info about RTI for lithics documentation. https://rtimage.us/?page_id=27.
Pétillon, JM. 2005. Tir expérimental de pointes à base fourchue en bois de renne. In: Dujardin, V (ed.), Industrie osseuse et parures du Solutréen au Magdalénien en Europe, 243–256. Paris: Société préhistorique française.
Pétillon, JM. 2016. Thirty Years of Experimental Research on the Breakage Patterns of Stone Age Osseous Points. Overview, Methodological Problems and Current Perspectives. In: Multidisciplinary approaches to the study of Stone Age weaponry, Ovita, R and Sano, K (eds.), 47–63. Springer. DOI: https://doi.org/10.1007/978-94-017-7602-8_4
Pokines, JT. 1998. Experimental Replication and Use of Cantabrian Lower Magdalenian antler projectile points. Journal of Archaeological Science, 25(9): 875–886. DOI: https://doi.org/10.1006/jasc.1997.0269
Reilly, P, Callery, S, Dawson, I and Gant, S. 2021. Provenance Illusions and Elusive Paradata: When Archaeology and Art/Archaeological Practice Meets the Phygital. Open Archaeology, 7(1): 454–481. DOI: https://doi.org/10.1515/opar-2020-0143
Rick, TC, Erlandson, JM and Vellanoweth, RL. 2001. Paleocoastal marine fishing on the Pacific coast of the Americas: Perspectives from Daisy Cave, California. American Antiquity, 66(4): 595–613. DOI: https://doi.org/10.2307/2694175
Roberts Thompson, AD and Williams, M. 2016. Reflectance transformation imaging (RTI) of a little known southeastern copper plate. Southeastern Archaeology, 35(1): 70–80. DOI: https://doi.org/10.1179/2168472315Y.0000000008
Selmo, D, Sturt, F, Miles, J, Basford, P, Malzbender, T, Martinez, K, Thompson, C, Earl, G and Bevan, G. 2017. Underwater reflectance transformation imaging: A technology for in situ underwater cultural heritage object-level recording. Journal of Electronic Imaging, 26(1): 011029. DOI: https://doi.org/10.1117/1.JEI.26.1.011029
Semenov, SA. 1964. Prehistoric technology: An experimental study of the oldest tools and artefacts from traces of manufacture and wear. Translated, and with a preface by MW Thompson. London: Cory, Adams & Mackay, 1964.
Soffer, O. 2004. Recovering Perishable Technologies through Use Wear on Tools: Preliminary Evidence for Upper Paleolithic Weaving and Net Making. Current Anthropology, 45(3): 407–413. DOI: https://doi.org/10.1086/420907
Soffer, O, Adovasio, JM and Hyland, DC. 2001. Perishable technologies and invisible people: nets, baskets, and “Venus” wear ca. 26,000 BP. Enduring Records: the Environmental and Cultural Heritage, 233–45.
Solem, D and Nau, E. 2020. Two New Ways of Documenting Miniature Incisions Using a Combination of Image-Based Modelling and Reflectance Transformation Imaging. Remote Sensing, 12(10): 1626. DOI: https://doi.org/10.3390/rs12101626
Stammers, RC, Harries, A, Spry, C, Armstrong, B and Caruana, M. 2017. Holocene LSA Archaeology from Equus Cave; Buxton-Norlim Limeworks, South Africa: An Analysis of the Bone Tool Assemblage. South African Archaeological Bulletin, 72(206): 103–115.
Stodiek, U. 1993. Zur Technologie der Jungpaläolithischen Speerschleuder: eine Studie auf der Basis Archäologischer, Ethnologischer und Experimenteller Erkenntnisse [Unpublished doctoral dissertation]. Tübingen.
Stone, EA. 2009. Wear on Magdalenian Bone Tools: A New Methodology for Studying Evidence of Fiber Industries. In: North European symposium for archaeological textiles X, Strand, EB, Gleba, M, Mannering, U and Munkholt, C (eds.), 225–232. Oxbow Books.
Stone, EA. 2013. The Role of Ethnographic Museum Collections in Understanding Bone Tools. In: From these bare bones: Raw materials and the study of worked osseous objects, Choyke, A and O’Connor, S (eds.), 25–37. Oxbow Books.
Struckmeyer, K. 2011. The bone tools from the dwelling mound Feddersen Wierde, Germany, and their functions. In: Written in bones: Studies on technological and social contexts of past faunal skeletal remains, 187–195. Institute of Archaeology, University of Wrocław.
Tourloukis, V, Thompson, N, Panagopoulou, E, Giusti, D, Konidaris, GE, Karkanas, P and Harvati, K. 2018. Lithic artifacts and bone tools from the Lower Palaeolithic site Marathousa 1, Megalopolis, Greece: Preliminary results. Quaternary International, 497: 47–64. DOI: https://doi.org/10.1016/j.quaint.2018.05.043
Tyzzer, EE. 1936. The “Simple bone point” of the shell-heaps of the northeastern Algonkian area and its probable significance. American Antiquity, 1(4): 261–279. DOI: https://doi.org/10.2307/275242
Van Gijn, A. 2005. A functional analysis of some Late Mesolithic bone and antler implements from the Dutch coastal zone. In: From hooves to horns, from mollusc to mammoth: Manufacture and use of bone artefacts from prehistoric times to the present: Proceedings of the 4th meeting of the ICAZ worked bone research group at Tallinn, 26th–31st of August 2003, 47–66. Ajaloo Instituut.
Watson, A and Gleason, M. 2016. A comparative assessment of texture analysis techniques applied to bone tool use-wear. Surface Topography: Metrology and Properties, 4: 1–18. DOI: https://doi.org/10.1088/2051-672X/4/2/024002
Wojtczak, D and Kerdy, M. 2018. They left traces. Preliminary analyses of micro-wear traces on bone and antler tools from Sutz-Lattrigen Aussen, Lake Bienne, Switzerland. Journal of Archaeological Science: Reports, 17: 798–808. DOI: https://doi.org/10.1016/j.jasrep.2017.12.038
Yellen, J, Brooks, A, Cornelissen, E, Mehlman, M and Stewart, K. 1995. A Middle Stone Age Worked Bone Industry from Katanda, Upper Semliki Valley, Zaire. Science, 268(5210): 553–556. DOI: https://doi.org/10.1126/science.7725100