Lithic microdebitage has great archaeological potential to elucidate ancient stone tool production. So far, archaeologists have collected soil samples, separated them into size fractions, and analyzed them manually under a microscope to identify microdebitage. This time- and labor-intensive process has limited the number of samples and introduced intra- and inter-observer errors. Here, we discuss lithic microdebitage analysis with a dynamic image particle analyzer. This machine takes videos of soil particles as they fall from a chute. Software tracks them and measures their dimensions. Since sieving is no longer necessary, microdebitage analysis proceeds more quickly and processes samples within a few minutes. The standardized output allows the objective analysis of lithic microdebitage. We compare the angularity of c. 120,000 particles in an archaeological soil sample with experimental microdebitage. While the distributions show intriguing overlaps, we conclude that the most angular archaeological particles are not microdebitage but reflect a software glitch.
Archaeologists tend to produce slow data that is contextually rich but often difficult to generalize. An example is the analysis of lithic microdebitage, or knapping debris, that is smaller than 6.3 mm (0.25 in.). So far, scholars have relied on manual approaches that are prone to intra- and interobserver errors. In the following, we present a machine learning–based alternative together with experimental archaeology and dynamic image analysis. We use a dynamic image particle analyzer to measure each particle in experimentally produced lithic microdebitage (N = 5,299) as well as an archaeological soil sample (N = 73,313). We have developed four machine learning models based on Naïve Bayes, glmnet (generalized linear regression), random forest, and XGBoost (“Extreme Gradient Boost[ing]”) algorithms. Hyperparameter tuning optimized each model. A random forest model performed best with a sensitivity of 83.5%. It misclassified only 28 or 0.9% of lithic microdebitage. XGBoost models reached a sensitivity of 67.3%, whereas Naïve Bayes and glmnet models stayed below 50%. Except for glmnet models, transparency proved to be the most critical variable to distinguish microdebitage. Our approach objectifies and standardizes microdebitage analysis. Machine learning allows studying much larger sample sizes. Algorithms differ, though, and a random forest model offers the best performance so far.
Dental mesowear analysis can classify the diets of extant herbivores into general categories such as grazers, mixed-feeders, and browsers by using the gross wear patterns found on individual teeth. This wear presumably results from both abrasion (food-on-tooth wear) and attrition (tooth-on-tooth wear) of individual teeth. Mesowear analyses on extinct ungulates have helped generate hypotheses regarding the dietary ecology of mammals across space and time, and recent developments have expanded the use of dental mesowear analysis to herbivorous marsupial taxa including kangaroos, wombats, possums, koalas, and relatives. However, the diet of some of the most ubiquitous kangaroos (e.g., Macropus giganteus) along with numerous other species cannot be successfully classified by dental mesowear analysis. Further, it is not well understood whether climate variables (including precipitation, relative humidity, and temperature) are correlated with dental mesowear variables including various measures of shape and relief. Here, we examine the relationship between dental mesowear variables (including traditional methods scoring the sharpest cusp and a new potential assessment of multiple cusps) and climate variables in the grazers/mixed feeders Macropus giganteus and Macropus fuliginosus, and the obligate browser Phascolarctos cinereus. We find that dental mesowear of mandibular teeth is capable of differentiating the dietary habits of koalas and the kangaroo species. Furthermore, both Macropus giganteus and Phascolarctos cinereus exhibit mesowear correlated with mean minimum temperature, while Macropus fuliginosus dental mesowear is unaffected by temperature, despite significant differences in mean minimum and mean maximum temperature across their distribution (and in the specimens examined here). Contrary to expectations that individuals from drier regions would have blunter and lower relief teeth, dental mesowear is unrelated to proxies of relative aridity—including mean annual precipitation and relative humidity. Collectively, dental mesowear in these marsupials is related to feeding behavior with increased wear in cooler regions (in Macropus giganteus and Phascolarctos cinereus) potentially related to more or different food resources consumed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.