Browse Articles

Expectation vs. reality: How stereotypes and expectation disconfirmation affect job evaluations in online labor markets

In online labor markets, reputation determines both job opportunities and pay and even small disparities can translate into significant economic differences. While gender-related bias in reputational metrics such as likes, reviews, and ratings has been empirically documented, the mechanisms through which they arise remain insufficiently understood. This study presents results from an online experiment, varying workers’ gender, domain, and performance, to investigate how expectations about workers are formed and how they affect subsequent evaluations. Drawing on Expectation Disconfirmation Theory and Role Congruity Theory, we test whether the effect of disconfirmation (i.e., discrepancies between expected and actual performance) on evaluations varies with congruity (i.e., stereotypical fit between gender and domain). Contrary to our hypothesis, workers in congruent settings do not elicit higher expectations than those in incongruent roles. Similarly, the effect of positive and negative disconfirmation does not vary with congruity. However, exploratory analyses suggest that congruity does affect expectations and evaluations when individuals hold strong gender-domain associations. In case of stereotypical associations, individuals expect higher performance of workers in congruent domains and evaluate them more leniently if they fail to meet expectations. Our findings contribute to understanding why identical performances might be judged differently depending on gender-domain associations, expectations, and disconfirmation.

Jumbo phage–mediated transduction of genomic islands

Bacteria acquire new genes by horizontal gene transfer, typically mediated by mobile genetic elements (MGEs). While plasmids, bacteriophages, and certain integrative and conjugative elements are well characterized, the broader diversity of MGEs remains poorly understood. Here, we cultured the bacterium Pseudomonas fluorescens SBW25 with sterile filtrate obtained from garden compost communities. Genome sequencing of derived colonies revealed acquisition of three different mobile elements, each integrated immediately downstream of tmRNA , each flanked by direct repeats, and each encoding a tyrosine integrase ( intY ) plus putative phage defense systems. Absent are genes with recognized roles in autonomous transfer. Interrogation of DNA sequence databases showed that similar elements are widespread in the genus Pseudomonas and beyond, with Vibrio Pathogenicity Island-1 from Vibrio cholerae as a notable example. Bioinformatic analyses reveal evidence of extensive horizontal transfer among diverse hosts. Detailed analysis of a single element, I55, showed that it is transferred between cells by jumbo phages, and confers fitness benefits via a type II restriction-modification system.

Associations of social determinants of health and patient safety in perinatal care: Protocol for a systematic review with meta-analysis

Background Associations between adverse social conditions and poor health are well documented – also in perinatal care. However, research into the actual ramifications of such disparities for perinatal patient safety remains inconclusive. Therefore, to achieve a comprehensive understanding of the risk and burden of patient harm to disadvantaged people, we aim to systematically review current evidence on social determinants of health (SDoH) and perinatal patient safety. Objective This study protocol outlines definitions, methods, and procedures for a systematic literature review with meta-analysis aiming to synthesise the research base on the associations between SDoH and patient safety in perinatal care. Methods Adhering to PRISMA guidelines, a literature search will be conducted for a systematic review in MEDLINE (PubMed), Scopus database, CINAHL (EBSCO), and Embase (Elsevier) for quantitative studies reporting associations between SDoH and patient safety measures in perinatal care. Data extraction will include study design, population, SDoH variables, outcome measures, effect sizes, and control variables. If deemed feasible after assessment of heterogeneity, narrative synthesis of findings will be complemented by conducting meta-analyses of pooled effect sizes. Methodological quality of included studies will be assessed using JBI Critical Appraisal Tools, and the certainty of evidence using the GRADE tool. This protocol is registered on PROSPERO (CRD420251090149) and OSF ( https://doi.org/10.17605/OSF.IO/UP3JS ). Discussion The review aligns well with current global efforts to promote safe perinatal care and presents an innovative, comprehensive approach for assessing the associations between SDoH and patient safety. The review will provide the first systematic synthesis of current evidence of SDoH and patient safety in perinatal care. Anticipated limitations include heterogeneity of study designs, measures, and outcomes, with expected predominance of observational studies, which may limit causal inferences. However, this review will provide a valuable foundation for further empirical research and interventions to enhance equitable and safe perinatal care.

Extreme triple oxygen isotope fractionation in <i>Equisetum</i>

Triple oxygen isotope values of xylem water were measured along the length of smooth horsetail stems ( Equisetum laevigatum ). Extreme isotope enrichment is observed moving from base to stem tip. δ 18 O values range from −8.3‰ at the base to 82.6‰ at the tip. Δ′ 17 O values range from 0 to −1,797 per meg. The δ 18 O and Δ′ 17 O values are the most extreme measured for any terrestrial material and expand the known range of Δ′ 17 O values by fivefold for mass-dependent fractionation on Earth. The extreme isotope enrichments are explained using a hybrid evaporation/chain-of-lakes model, allowing us to refine the leaf respiration coefficient to θ k = 0.511 ± 0.001. This new value is required to explain the low Δ′ 17 O values previously measured in desert plants and animals and is critical when using fossil samples for paleoclimate reconstruction. Coexisting phytoliths and stem water were also measured. The 1000 ln 18 α silica-water value at the plant base (35.89‰) appears to be in isotopic equilibrium, with far smaller fractionations of 10.3‰ near the tip. The smaller fractionations at higher levels are explained by continual silica deposition as the plant elongates and the δ 18 O values of each segment become higher. The overall integrated phytolith value is a combination of early and late silica growth. The Δ′ 17 O silica − Δ′ 17 O water values are not in equilibrium, explained by a kinetic isotope effect, with a λ value of 0.5205 vs. 0.5244 for equilibrium. Phytolith isotope values may lead to erroneous interpretations for paleoclimate reconstruction.

The impact of ESG performance on corporate innovation—Empirical evidence from Chinese pharmaceutical listed companies

Environmental, social and governance performance is an important driver for companies to achieve innovative and high-quality development, and the relationship between sustainable development and corporate innovation is one of the key concerns of pharmaceutical companies. This study aims to reveal the intrinsic role mechanism of sustainable transformation of pharmaceutical enterprises in high-quality innovation development. Based on the relevant data of China’s A-share pharmaceutical listed companies from 2015 to 2022, the study establishes a fixed effect model and a mediation effect model, empirically examines the impact of ESG performance on the innovation of pharmaceutical listed companies, and deeply explores the role mechanism of R&amp;D personnel investment and government subsidies in it. The study finds that: (1) ESG performance significantly enhances firms’ innovation inputs and green innovation outputs; (2) this facilitating effect is more substantial among non-state-owned firms and firms in the eastern region; (3) good ESG performance has a more pronounced facilitating effect on innovation inputs of large firms, while it has a more prominent effect on the green innovation outputs of small and medium-sized firms; and (4) the mechanism analysis indicates that internal R&amp;D personnel investment and external government subsidies play a significant mediating effect between ESG performance and corporate innovation. This study provides new empirical evidence for understanding the mechanism of ESG performance on corporate innovation in Chinese pharmaceutical companies. It provides theoretical support for proactively promoting the sustainable transformation and deepening the development of green innovation in Chinese pharmaceutical companies. In addition, there are important policy implications for promoting the sustainable transformation of pharmaceutical companies: the establishment of differentiated ESG incentive policies, the improvement of R&amp;D talent training mechanisms, and the optimization of government subsidy allocations, focusing on supporting green innovation by companies with excellent ESG performance.

Monocyte-derived exosomal periostin driven by histone lactylation contributes to retinal neovascularization

Exosomes are signaling messengers facilitating intercellular communication by delivering molecular cargo. Here, we observed increased levels of exosomal periostin (POSTN) in the blood plasma of patients with proliferative diabetic retinopathy (PDR). Monocytes contributed to elevated levels of exosomal POSTN, and deletion of Postn in myeloid cells reduced retinal neovascularization (RNV). Monocytes isolated from the blood of PDR patients or under high glucose conditions exhibited heightened glycolytic activity and elevated histone lactylation levels, particularly H4K8 lactylation (H4K8la). Knockout of hexokinase 2 ( Hk2 ) in myeloid cells led to reduced H4K8la and POSTN levels and inhibited RNV. Exogenous exosomal POSTN partially reversed the angiogenic defects caused by Postn or Hk2 deletion in myeloid cells. Mechanistically, exosomal POSTN stabilized hypoxia-inducible factor -1 alpha and upregulated the expression of angiogenic genes. Notably, treatment with metformin reduced RNV by decreasing monocyte glycolysis and lowering exosomal POSTN levels. In summary, these findings underscore the critical role of circulating exosomal POSTN in RNV and highlight its potential as a therapeutic target for angiogenic retinopathies.

Unveiling and verification of mitochondria-related genes as potential diagnostic biomarkers in ulcerative colitis based on bioinformatics analysis and experimental validation

Immune dysregulation is a pathogenic factor in ulcerative colitis (UC), in which mitochondrial involvement is increasingly recognized. This study aimed to construct a diagnostic model using mitochondria-related genes, identify new target genes, and illuminate the roles of mitochondria-related genes in energy metabolism, immune infiltration and the pathogenesis of UC. RNA expression data from 465 patients with UC and 154 healthy controls (HCs) were obtained from the GEO database. A total of 128 mitochondria-related differentially expressed genes (Mito-DEGs) were identified between patients with UC and HCs. A diagnostic model constructed from 20 genes showed satisfactory discrimination, calibration functions, clinical benefits and clinical impacts. Enrichment and immune infiltration analyses revealed significant differences in mitochondrial structure and function, immune cell disorders, and signaling pathway activation between the high- and low-mitochondrial gene expression UC groups. Correlations between mitochondrial structure and function and immune cells were evaluated. Single-cell RNA sequencing data were used to analyze the hub gene distribution, cell‒cell communication, and enrichment. Cell‒cell communication analysis revealed that immune response and pathogenesis pathways are activated in UC. Experiments revealed that the expression of the ACADM and ACADSB genes was decreased in UC patients. Mitochondrial dysfunction contributes to the pathogenesis of UC by altering energy metabolism, promoting immune disorders and activating pathogenic signaling pathways. The mitochondria-related genes are valuable for the diagnosis of UC. ACADM and ACADSB may play important roles in UC pathogenesis.

Male-derived PBP4 is essential for sperm competition by mediating sperm motility in moths

Male moths rely on antennal olfactory sensilla to detect female-emitted sex pheromones during mate localization, a process critically mediated by pheromone-binding proteins (PBPs) that facilitate ligand transport to dendritic membrane-bound pheromone receptors. Intriguingly, a divergent PBP paralog (PBP4) in noctuid moths has been identified exhibiting predominant expression in the male reproductive system, though its functional significance remained elusive. In this study with Spodoptera exigua , we demonstrated that PBP4 is specifically expressed in male accessory glands (MAGs) and transferred to females during copulation. CRISPR-mediated PBP4 knockout (PBP4 −/− ) significantly reduced male reproductive success via impaired sperm competition, evidenced by diminished offspring sired. Transcriptomic and qPCR analyses revealed expressional correlation between PBP4 and the odorant receptor coreceptor (Orco), and similar sperm competition deficits were observed in Orco −/− males. Furthermore, both mutants exhibited compromised sperm motility and reduced sperm transfer from spermatophores to spermathecae. In vitro sperm motility assays showed that a specific agonist of Orco, VUAA1, robustly enhanced sperm motility in an Orco-dependent manner. Analyses of four additional moth species, including noctuids and non-noctuids, revealed a similar specific expression in the male reproductive tracts in all of them. Our findings unveil a noncanonical role for PBP4 in reproductive fitness and evolutionary conservation of the PBP-OR signaling pathway.

Trophic niche variation across the pan-Arctic coastal continuum

We analyzed stable carbon and nitrogen isotope values (δ 13 C and δ 15 N, respectively) for pan-Arctic coastal primary producers and consumers to detect large-scale regional trends both temporally and spatially. To facilitate comparison, we grouped coastal habitats into fjords, lagoons, shelves, and straits as four “coastscapes”. We gathered over 12,000 rows of data collected over 24 years (between 1999 and 2022) from 34 different field campaigns across the coastal Arctic (63 to 81°N and 177°W to 33°E). Our goal was to examine the isotopic patterns in pelagic and sediment particulate organic matter (pPOM and sPOM, respectively) and four consumer groups (deposit feeders, opportunists/scavengers, predators, and suspension feeders) among the four coastscapes. We found that despite the enormous spatial range of data, both pPOM and sPOM became 2.1‰ and 2.2‰ more 13 C-depleted per decade, respectively, with parallel decreases in the δ 13 C values in consumers. The significant decrease is likely attributed to the increased contributions of 13 C-depleted terrestrial organic matter across the Arctic coasts from freshwater inputs and coastal erosion in concert with diminishing sea ice that supports sympagic microalgae. Across all Arctic coastscapes, consumer groups exhibited overlapping isotopic composition, notably with wide δ 13 C ranges that indicated assimilation of multiple organic matter sources, including terrestrial organic matter, organic matter derived from marine phytoplankton and sea ice algae, macroalgae, and potentially benthic microalgae or degraded organic matter. This consistent pattern across coastscapes provides evidence of the trophic plasticity possessed by Arctic consumers, how coastal food webs respond to climate warming, and the signature of terrestrialization imprinted on the pan-Arctic coastal isoscape.

Serum response factor is essential for endometrial function and prevention of inflammatory fibrosis

Pregnancy requires a supportive uterine environment facilitated by steroid hormone–regulated differentiation of endometrial stromal fibroblasts into decidual cells and tight control of inflammation. Serum response factor (SRF) is a widely expressed transcription factor essential for mesenchymal cell growth and differentiation with noted roles in hormonal regulation of muscle tissues but little characterization in reproductive organs. Here, we reveal that endometrial SRF is dysregulated in human endometriosis and is critical for female reproductive success in mice through regulation of endometrial stromal and epithelial cells. Immunohistochemical analysis identified decreased endometrial SRF expression in infertile endometriosis patient tissues. RNAi-based SRF knockdown in human endometrial stromal cells resulted in disrupted cytoskeletal structure, viability, and decidual response. Conditional Srf knockout mice (Srf d/d ) generated using the Pgr Cre were infertile with implantation failure, lack of decidual response, and development of severe endometrial fibrosis. Single-cell RNA sequencing identified dramatic changes in Srf d/d endometrial cell populations including large-scale myeloid immune cell infiltration. Srf d/d stromal fibroblasts displayed aberrant cytoskeletal and extracellular matrix gene expression and downregulation of genes important for decidual growth response. Srf -deficient epithelial cells displayed the most prominent dysregulation, strongly overexpressing estrogenic innate inflammatory genes including C3 and Lcn2 and fibrogenic genes Mmp7 and Fbln1 , all of which paralleled patterns in human endometriosis patient data that we identified through comparative analysis using a published single-cell atlas. These results demonstrate the profound impact of SRF on endometrial homeostasis with relevance to human endometriosis–related infertility.

Correction: Memory improving effect of silkworm larva on insulin resistance related cognitive impairment model

Structural pores not required: Antimicrobial peptides induce ion permeabilization of lipid membranes through transient water channels

Antimicrobial peptides (AMPs) are highly potent and broad-spectrum antibiotics, found as components of the innate immune system in almost all forms of life. Despite being commonly accepted that the mechanism of action of AMPs is associated with the permeabilization of the cell membrane, the structural and dynamical means by which this occurs are still heavily debated. In this work, we employ experimental time-resolved small angle X-ray scattering to follow in real time the AMP-induced ion transport in lipid vesicles, while simultaneously resolving the membrane structure and peptide partitioning. For several natural AMPs, we show that they can effectively permeabilize the lipid membrane despite only binding peripherally to the outer membrane leaflet. Our experiments reveal rapid ion transport associated with AMP binding, yielding salt equilibration in a few tens of milliseconds, while not detecting evidence of transversal and structurally stable peptide pores. On the contrary, new analysis of previously reported all-atom molecular dynamics simulations shows that lipid flip-flop, accelerated by the peripherally bound peptides, leads to the formation of transient, ion-conducting water channels. A corresponding diffusional model indicates that such short-lived, transient pores explain the observed ion transport better than stable pores.

Application of deep learning in behavior recognition and early warning system for campus safety management

Campus safety is an essential concern as schools, colleges, and universities work to create secure environments for students, staff, and visitors. Many existing security systems are not fully effective at detecting unusual behaviors or sending fast alerts, which can delay responses to potential threats. To improve this, the research introduces DeepCARE(Deep-learning-based Campus Anomaly &amp; Risk Evaluation), a deep learning-based framework designed to enhance behavior recognition and early warning for campus security. DeepCARE combines convolutional neural networks (CNNs) with long short-term memory (LSTM) networks to process real-time video footage and detect abnormal activities such as aggression, unauthorized access, and people staying too long in restricted areas. The system’s main feature is its hybrid model, where CNNs extract key visual features from surveillance footage while LSTM networks analyze these features over time to recognize behavior patterns. DeepCARE also includes an anomaly detection module using autoencoders, which helps improve the system’s accuracy and reduces false alarms. This makes DeepCARE a flexible and scalable solution, suitable not only for educational campuses but also for public spaces, transport hubs, and smart cities. By applying deep learning, DeepCARE supports early risk detection and faster response times, helping security teams create safer spaces. Experimental results show that DeepCARE achieves a behavior recognition accuracy of 94.5%, performs 8% better than traditional methods, and shortens emergency response times by 30%.

Homology-mediated transformation of frog-killing fungus <i>Batrachochytrium dendrobatidis</i> illuminates chytrid development and pathogenesis

The chytrid fungus Batrachochytrium dendrobatidis ( Bd ) infects amphibians and causes chytridiomycosis, a disease linked to global amphibian decline. Despite its ecological importance, Bd has lacked robust tools for genetic manipulation, limiting molecular insights into its development and pathogenicity. Here, we establish homologous recombination as a system for stable transformation of Bd , enabling targeted chromosomal integration of exogenous DNA, endogenous protein tagging, and targeted gene deletion. We use this system to visualize Bd infection in live amphibians to enhance understanding of host invasion and pathogenesis. We also use this system to test a previous hypothesis regarding the role of chitin synthases in Bd development by tagging the endogenous chitin synthase Myo17D and observing its rapid relocalization to the plasma membrane during de novo cell wall assembly. Finally, we use our homologous recombination approach for targeted gene deletion by knocking out the URA3 locus, and confirm the resulting genotype and phenotype via sequencing and drug resistance assays. This genetic transformation system offers a foundational tool for molecular studies of Bd , advancing our capacity to dissect molecular mechanisms of chytrid pathogenesis.

Attention to detail: A conditional multi-head transformer for traffic sign recognition

The challenge of traffic sign detection and recognition for driving vehicles has become more critical with recent advances in autonomous and assisted driving technologies. Although object recognition problems, particularly traffic sign recognition, have been extensively studied, most Vision Transformer (ViT) models still rely on static attention mechanisms with fixed projection matrices (Q, K, and V). Using this mechanism limits the ViTs to handle real-world problems such as object detection and traffic sign recognition, etc. Problems, such as partially or fully obscured signs, changes in illumination, and weather conditions, result in subpar feature extraction, which compounds the misclassification problem. To overcome this challenge, a Conditional Visual Transformer (CViT) is proposed in this research, which dynamically adapts feature aggregation, Q, K, and V projections, as well as attention-based mechanisms, based on the input sign type. Its main component consists of a controlled failure deep learning model using a CViT that targets specific types of traffic signs through varying feature extraction and attention adjustments, resulting in high classification performance and minimizing misclassifications. Furthermore, an adaptive gating technique is employed that optimally adjusts the projection matrix across different traffic signs. The proposed CViT achieved an overall accuracy of 99.87%, with a Micro Precision of 99.07%, a Macro Recall of 94.3%, and a Macro F1 Score of 99.07%, respectively. These results demonstrate the potential of CViT to improve both the efficiency and reliability of traffic sign recognition in autonomous driving applications.

A functional clock in only two dorsal clock neurons is sufficient to restore the basal circadian activity pattern of <i>Drosophila melanogaster</i>

Circadian clocks form complex networks to orchestrate the behavior and physiology of animals. Elucidating the organization of these clock networks is critical to understanding how circadian clocks achieve robust timing. Clock neurons have been best characterized in the model organism Drosophila melanogaster . At the early stages of development, Drosophila has a clock network consisting of only nine neurons per hemisphere. This set of clock neurons consists of lateral and dorsal neurons and persists in the adult fly, in which the clock network is greatly expanded to a total of ~240 neurons. The function of the early-born lateral clock neurons is well established in the adult fly. They control morning and evening activity and are responsible for circadian rhythmicity under constant conditions. Only recent studies have begun to dissect the function of the dorsal clock neurons. Using clock rescue experiments, we show here that a functional clock in only four of them, the DN 1a , is sufficient to restore a bimodal locomotor activity pattern and rhythmic activity under constant conditions in the adult fly. This is achieved by CCHamide-1 signaling most likely to the lateral morning neurons, which are considered the main pacemakers of the clock, and by glutamate signaling to the lateral evening neurons. Interestingly, all the neurons involved are part of the early-born clock network, strongly suggesting that a core clock is already formed during early development. This set of clock neurons appears sufficient in the adult fly to drive the basal circadian activity pattern.

Self-management of type 2 diabetes using a co-designed text based mobile health (mHealth) intervention in Nepal: A study protocol for randomised controlled trial

Mobile health (mHealth) is increasingly being used for managing chronic diseases proving its feasibility, acceptability and effectiveness in improving the health outcomes. However, effectiveness of mHealth intervention for self-management of Type 2 diabetes remains unexplored in the context of Nepal. This study aims to develop and evaluate the effectiveness of a co-designed text based mobile health intervention for self-management of Type 2 diabetes in Nepal. We will conduct a six month, two arm (1:1) parallel group randomised control trial involving 154 adults with Type 2 diabetes. The primary outcome includes change in haemoglobin A1c level at six months follow up. Secondary outcomes include change in self-management outcomes such as self-efficacy, perceived support, diabetes related stress, health related quality of life and self-care activities. Primary and secondary outcomes will be summarized descriptively, and comparisons will be made using chi-square tests for categorical data and independent sample t-tests for continuous data. The trial is registered at ClinicalTrials.gov (NCT06623006).

Micro-offline gains do not reflect offline learning during early motor skill acquisition in humans

While practicing a new motor skill, resting for a few seconds can improve performance immediately after the rest. This improvement, referred to as “micro-offline gains” (MOGs), has been interpreted as rapid offline learning and is thought to be supported by neural replay of the trained movement sequence during rest. Here, we provide evidence that MOGs reflect only transient performance benefits, partially mediated by motor planning, and not offline learning. In five experiments, participants trained to produce a sequence of finger movements as many times as possible. When participants trained during 10-s practice periods, interleaved with 10-s rest periods, they produced more correct keypresses during training than participants who trained continuously without taking breaks. However, this benefit vanished within seconds after training, when both groups performed under comparable conditions, revealing similar levels of skill acquisition. This challenges the idea that MOGs reflect offline learning, which, if present, should result in sustained performance benefits, compared to training without breaks. Furthermore, we observed persistent MOGs even when participants produced random, nonrepeating sequences, indicating that MOGs do not relate to the offline consolidation of a sequence-specific memory. Rather, sequence-specific learning was only evident online, during practice. Finally, MOGs were diminished when participants could not preplan the first few movements of an upcoming practice period. Our results suggest that MOG are simply an effect of slowing during practice, combined with the possibility of preplanning the initial movements after rest, and therefore do not serve as a reliable metric for offline learning or skill-related consolidation.

Machine learning models for risk prediction of age-related macular degeneration in Fujian eye study

Objective Age-related macular degeneration (AMD) is a retinal disorder that significantly impairs vision. This study investigates various machine learning models for predicting AMD risk, laying the groundwork for further research using big data and determining the most effective predictive model. Methods Utilizing data from 8211 records with 39 features from the Fujian Eye Study, a cross-sectional epidemiological investigation, several machine learning models were developed and assessed. The models included logistic regression (LR), K-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT), random forest (RF), light gradient boosting machine (LightGBM), and extreme gradient boosting (XGBoost). Data preprocessing, feature selection, and model training were all key components of the study. Results After evaluating multiple models, the logistic regression model emerged as the most accurate, achieving a balanced accuracy of 0.6364. Among the predictive features, educational background had the highest influence on the model’s predictions, with an average SHAP (SHapley Additive exPlanations) value of 0.8199. Other significant factors included outdoor time and left eye spherical equivalent (OSSE), with SHAP values of 0.6474 and 0.6377, respectively. Conclusion This study confirms that logistic regression is the most effective machine learning model for predicting AMD risk, with educational background identified as the most critical risk factor.

Miocene and Pliocene ice and air from the Allan Hills blue ice area, East Antarctica

Antarctic ice cores provide a unique archive of Earth’s atmosphere and its largest extant ice sheet. The oldest continuous ice core extends back 800 ky, though discontinuous ice cores from the Allan Hills blue ice area (BIA) have been shown to preserve snapshots of ice and air back to at least 2.7 million years ago (Ma). Here, we provide snapshots of putatively Miocene and Pliocene ice and air from shallow ice cores drilled in the Allan Hills BIA. The ice, dated using the deficit in 40 Ar in ancient air compared to the modern atmosphere, is stratigraphically complex. Nevertheless, surface temperatures inferred from water isotopes correlate with sample age and indicate 12 ± 2 °C of cooling in Antarctica between 6 Ma and the late Pleistocene. Basal ice is nearly devoid of gases and remains to be dated with existing methods. This undated ice is characterized by an isotopic temperature 5 ± 1 °C warmer than the oldest dated (6 million year old) sample. We speculate that this ice reflects surface snowpack or permafrost that was preserved by the growth of the East Antarctic ice sheet in the Middle to Late Miocene.