In our paper ‘Transport of animals underpinned ritual feasting at the onset of the Neolithic in southwestern Asia’, we used a suite of analytical techniques to examine the geographical origins of wild boars that were used for prehistoric feasting activities at the Early Neolithic archaeological site of Asiab, western Iran.
The techniques included established histological methods for analysing microscopic growth patterns inside the wild boar’s teeth and three geochemical proxies:
- Tooth enamel phosphate stable oxygen isotope values of measured using a Sensitive High Resolution Ion Microprobe (SHRIMP)
- Tooth enamel strontium isotope ratios measured using a Laser Ablation Multi Collector Inductively Coupled Plasma Mass Spectrometer (LA–MC–ICP–MS)
- Barium trace element maps of tooth enamel and dentine measured using a Laser Ablation Inductively Coupled Plasma Mass Spectrometer (LA–ICP–MS).
While the SHRIMP has previously been used to analyse teeth from humans and non-human primates (e.g., Green et al. 2022), this is the first time that it was used on zooarchaeological material to address questions of past human–animal interactions. For this reason, we were not just interested in collecting data that would tell an exciting story. We wanted to bring as much transparency as possible to the method so that others could reproduce our procedures and re-use our data in future analyses.
Taking extra steps to making SHRIMP analyses of teeth reproducible
The SHRIMP allows us to target ~20 µm wide analytical spots on the surface of a polished thin section. This provides us with a high-level of control over what is analysed. In our study, we placed analytical spots as close as possible to the enamel dentine junction (EDJ) and spaced them at intervals corresponding to approximately one week of enamel secretion. Although the composition of enamel laid down during enamel secretion is later attenuated during enamel maturation (see Green et al. 2025), this sampling strategy provides us with the most temporally resolved isotopic sequence that can be measured.
For this spot placement approach to be reproducible, we shared all the raw data collected during histological analysis (Supplementary Data 1). The file includes the individual enamel segments that were identified under the microscope, the length of each segment along the enamel dentine junction, how many days of secretion each segment represents, and the calculated Enamel Extension Rate, in µm per day. With a table like this, future analysis can be adjusted to whichever temporal resolution best suits a particular research question (i.e., analytical spots do not always have to be placed seven days apart).
One challenge with spot placement is aiming to get as close as possible to the enamel dentine junction without touching the dentine. When a spot touches the dentine, its measurement becomes compromised because it no longer reflects only the isotopic composition of enamel. And because dentine does not preserve as well as enamel does archaeologically, its composition may be diagenetically altered.
Although spot placement can be very precise as far as the ion beam is concerned, visibility of the surface of the sample can pose a challenge with placing analytical spots as close as possible to the enamel dentine junction without touching it. When the visibility is good (for example, in left-hand panel in the figure below), the enamel dentine junction presents as a thin line, and it is easy to tell where the enamel ends. However, when the visibility is poor (for example, in the right-hand panel in the figure below), a shadow may be cast over the enamel dentine junction and it is more difficult to determine where the adjacent enamel ends, and where the dentine starts.
Figure 1. Two examples of photographs of the sample surface, taken using the camera inside the Sensitive High Resolution Ion Microprobe (SHRIMP). Left: when visibility is good and the enamel dentine junction (EDJ) appears as a clear line. Right: when visibility is poor and the EDJ is cast in shadow.
To check the reliability of the measurements, it is important to inspect the photographs of the analytical spots taken inside the instrument and identify measurements that may have touched the EDJ. This process involves a certain level of subjectivity, as different people may reach a different decision.
To be fully transparent about how we assessed the reliability of our measurements, we created a SHRIMP Reproducibility Index, where others can inspect the placement of all of our analytical spots. The Index is available on the Open Access Framework profile for this study (https://osf.io/ 6fyeu/). It contains all the photographs taken inside the instrument and identifies which measurements were deemed unreliable. So that others can re-do our assessment and analysis, all raw measurements (including those deemed unreliable and not included in the final analysis) were included in Supplement Data 2.
Figure 2 Excerpt from the SHRIMP Reproducibility Index showing the placement and assessment of the reliability of all measurements. The full index can be accessed at https://osf.io/6fyeu/.
Advocating for Open Science and statistical reform
I (Petra Vaiglova) spend a lot of time in class talking with my students about the importance of Open Science and the need for statistical reform. We discuss why there is a reproducibility crisis, how we can use estimation thinking to avoid the dangers of Null Hypothesis Significance Testing, and what resources exist for helping us get started with embracing Open Science principles. For a full discussion, see my paper ‘How can we improve statistical training in archaeological science’ (2025, Journal of Archaeological Science).
How does this paper demonstrate that I practice what I preach? Firstly, we used estimation thinking to report effect sizes and their relevant confidence intervals. Specifically, when comparing the difference between the stable oxygen isotope values of sample ASB449 and samples ASB133–402, we spelled out the actual difference:
(mean of ASB133–402, n =167) – (mean of ASB449, n =88) = 2.8‰, 95% CI [2.5, 3.1]
This estimate tells us that most likely, the true mean between the two groups is captured by the interval 2.5 ‰ – 3.1 ‰. This difference is meaningful and can be interpreted, regardless of what the result of a null hypothesis significance test shows.
Secondly, we clarified what kind of interpretations we can draw with a sample size of five teeth. Although we collected 165 stable oxygen isotope values and 107 strontium isotope ratios, the dataset still only represents five wild boar individuals and the interpretations are thus constrained to identifying, rather than quantifying, the phenomenon of animal transport for the purposes of ceremonial feasting.
Thirdly, because humans make errors, we asked an independent researcher to double check our raw data files against our submitted supplementary files to make sure that no error was propagated during formatting.
Lastly, we took extra steps to make sure that our datasets and protocols could be reused by others. In addition to the supplementary materials published along with the paper—which include all the raw data (Supplementary Data 1–3) and the code used to create our figures (Supplementary Software 1–3)—we created a profile on the Open Science Framework (https://osf.io/6fyeu/) that contains additional materials others can use to build familiarity with the procedure. This includes high-resolution microscope images of the samples before and after analysis. Others can use these image to inspect in closer detail what our sampling strategy looked like and adapt similar approaches in future research.
Using illustration to further increase accessibility to the research (written by Kathryn Killackey)
Visualization is a powerful tool for synthesizing archaeological data and sharing interpretations with multiples audiences. It is also a complex process that pushes researchers to grapple with uncertainty and the fragmentary nature of archaeological data. In order to create an illustration that best supported this paper’s central argument, we first had to select a moment in time to represent one of the possible scenarios that led to long-distance transport of wild boar to Asiab. We considered several scenes from the different scenarios outlined in the paper, from a boar hunt in progress to the crania being deposited. We ultimately decided that showing different groups arriving with their boar contributions and feast preparation underway best visualized the key points of the paper, animal transport and community cooperation.
Figure 3. Artist’s depiction of the results of our study. Drawn by Kathryn Killackey.
Once we had agreed upon the illustration’s subject, I (Kathryn Killackey, a science illustrator) created a series of preliminary sketches. These sketches became the basis for us to hash out the finer details of the scene. This discussion raised new questions, such as how long would a field-dressed boar stay fresh and could a group transport it over distances of around 70 km on foot before it spoiled? What did the communal structure look like and how far above ground did it stand? What would the site’s inhabitants and guests wear on an important occasion? We answered these and more questions by drawing on the Asiab’s archaeological record when possible, as well as analogy with contemporary sites and ethnographic data. While we cannot be certain of many of the illustration’s details, its creation and dissemination helps us more fully envision a possible community event at Asiab that is both grounded in data and engaging for multiple audiences.