Sloths (Folivora) display remarkable morphological diversity yet share a highly conserved dental formula, whose occasional deviations offer key insights into their evolutionary history. Among nothrotheriids, Neogene species retained upper and lower caniniforms, whereas these teeth were lost in Pleistocene genera. Here we describe, for the first time, a poorly developed alveolus interpreted as an atavistic upper caniniform alveolus in an individual of Nothrotherium maquinense from the Late Pleistocene of Bahia, Brazil. Computed tomography and X-ray imaging confirm its internal anatomy and rule out taphonomic or pathological origins. The alveolus occupies the precise position of the caniniform in Neogene nothrotheriids and exhibits similar morphology, suggesting a partial re-expression of a lost ancestral trait. Beyond its developmental implications, this observation may offer additional context for interpreting the taxonomy of Nothropus, traditionally defined based on the presence of a lower caniniform alveolus. The occurrence of similar atavistic structures in related taxa indicates that this feature alone may be insufficient for diagnosis and warrants reevaluation within a broader evolutionary framework.
Environmental DNA (eDNA) metabarcoding is a powerful technique for assessing biodiversity in threatened Neotropical aquatic environments, yet the effect of water-level fluctuations in lentic systems on eDNA detectability and distribution remains poorly understood. Thus, we conducted a multi-year eDNA metabarcoding survey of the fish community in Ingleses Lake, a small urban Neotropical reservoir, across periods of high and low water levels. We found that eDNA detectability is shaped by hydrology-induced changes, since the mean ɑ-diversity per site was higher and the mean β-diversity was lower during low-water conditions. While fish assemblages maintained a stable core community (PERMANOVA F = 2.26, P = 0.056), a significant difference in the homogeneity of multivariate dispersions was detected (PERMDISP F = 22.23, P = 0.003). These results support the hypothesis that higher water-level conditions of the reservoir led to lower species detectability, probably due to increased eDNA dilution. Critically, combining data from campaigns of both high and low water yielded a near-asymptotic species accumulation curve, demonstrating that temporal replication across dynamic hydrological cycles is crucial for maximising eDNA detection efficiency in Neotropical lentic systems.
Forensic identification in advanced decomposition body is challenging when traditional methods become unreliable. This case report demonstrates the utility of 3D modeling of the frontal sinus for human identification in forensic anthropology and odontology. A decomposed body found in southern Brazil was successfully identified using 3D-2D superimposition technique, which overlays a three-dimensional model derived from antemortem computed tomography (CT) data onto postmortem radiographs. The frontal sinus, with its unique morphological patterns, served as a reliable anatomical marker for comparison. This cost-effective approach utilizing open-source software addresses the need for accessible identification methods in resource-limited forensic settings. The technique demonstrates superior accuracy compared to traditional 2D radiographic methods while remaining practical for institutes with limited imaging technology. Our findings supporteas the application of 3D-2D superimposition and highlight the importance of forensic experts training in 3D forensic imaging technology for enhanced human identification capabilities.
This article examines the struggles of Brazilian voice actors in response to the expansion of generative artificial intelligence (genAI) in the screen industries, and what these struggles reveal about worker-led AI governance beyond paradigmatic cases such as Hollywood. While existing scholarship has focused primarily on individual impacts of genAI on cultural workers or on policy debates at a macro level, far less attention has been paid to collective organizing, mobilization, and bottom-up forms of AI governance, particularly in Majority World contexts. Drawing on the power resources approach, the article analyzes how Brazilian voice actors mobilize institutional, associational, and discursive power in a sector historically shaped by weak unionization, informality, and global dependency within AI and screen value chains. Empirically, the study combines trade press analysis, policy documents, social media content analysis, and interviews with organizers related to the movement Dublagem Viva, the main association in the sector. The findings show that worker-led AI governance in this case is less about direct control over workplace deployment or sectoral bargaining, and more about reconfiguring the public and regulatory debate around AI through policy advocacy, coalition-building, and affective engagement with audiences.
Lithology classification based on well log data has attracted the attention of geologists, as the extraction of rock samples from the wellbore is a costly process. Although some machine learning techniques have recently been applied to the problem, the existing studies are not standardized in terms of data processing and evaluation protocols. This paper aims to bridge this gap by providing common ground for the comparison of prospective works on lithology classification from well logs. The proposed benchmark is inspired by a thorough review of the literature, from which the main aspects of an experimental protocol are captured under the principles of simplicity, variety, impartiality, and meaningfulness. The public datasets used in the experiments are the FORCE and Geolink repositories. Ten methods are tested, including six deep learning models and four shallow machine learning algorithms, and compared by nine distinct performance metrics under cross-validation. Different sequence lengths are tested as input for deep learning models to assess their contextual understanding capabilities. The experimental results using the proposed setup show that context-insensitive models outperform most context-sensitive ones, but with high correlation. The study also reveals that the problem of lithology classification is still not solved. Therefore, the benchmark presented herein serves as a fair protocol to assess different methods and can be used by future works as a simple and unbiased way to compare new approaches in lithology classification.
