Phthalate esters are among the most widely produced plasticizer families in the world. They are added to polyvinyl chloride and countless consumer goods—food packaging, cosmetics, personal care products, medical devices, floor coverings, and nail polish—to make rigid plastics flexible and durable. Because phthalates are not chemically bound to the polymer matrix, they leach, migrate, and evaporate into food, dust, water, and air. Decades of toxicological work have classified several of them as endocrine-disrupting chemicals, capable of interfering with hormone signaling, reproductive development, and metabolic regulation. Regulatory agencies in the European Union and the United States have restricted a series of the most concerning compounds, including DEHP, DBP, BBP, and DiBP, particularly in toys and childcare articles, prompting industry to pivot toward alternative plasticizers such as terephthalates and adipates.

That pivot is precisely what makes the new data so striking. The most prevalent plasticizer detected across all four maternal and fetal matrices was DEHT, or di(2-ethylhexyl) terephthalate, a structural isomer of the restricted DEHP that is widely promoted as a safer replacement. DEHT recorded the highest geometric mean concentration in urine at 403.7 micrograms per kilogram, followed by placenta at 318.6 micrograms per kilogram, blood at 304.5 micrograms per kilogram, and amniotic fluid at 275.0 micrograms per kilogram. In other words, a chemical introduced to reduce risk is now the dominant plasticizer burden in pregnant women, and it is not staying in the mother. Its presence in amniotic fluid—the liquid bathing the fetus—demonstrates that substitution has not eliminated fetal exposure, it has changed the chemical profile of that exposure.

Among the classical phthalates, the short-chain compounds told their own story. Diisobutyl phthalate (DiBP) showed the highest concentrations in placental tissue, at 195.8 micrograms per kilogram, and in urine, at 226.0 micrograms per kilogram, indicating both substantial body burden and efficient renal clearance. Dimethyl phthalate (DMP) and diethyl phthalate (DEP), lower-molecular-weight esters commonly found in cosmetics, fragrances, and personal care formulations, were mainly detected in amniotic fluid, at 128.2 and 64.8 micrograms per kilogram respectively. The distribution pattern is chemically meaningful: smaller, more water-soluble molecules tend to partition differently across biological membranes than bulkier, lipophilic ones, and their accumulation in amniotic fluid suggests they can traverse the placental barrier and equilibrate within the fetal compartment.

The placenta is often described as a protective shield, but pharmacologically it behaves more like a semi-permeable membrane governed by physicochemical rules. Molecular size, lipophilicity, ionization state, and protein binding all determine whether a compound crosses by passive diffusion, facilitated transport, or is retained. Plasticizers, being moderately lipophilic and of low molecular weight, are well positioned to cross. The Tunisian team’s measurements confirm this: substantial amounts of both phthalate and non-phthalate plasticizers were found on the fetal side of the barrier. Although the renal system contributes significantly to elimination—reflected in the high urinary concentrations of DEHT and DiBP—excretion is evidently not fast enough to prevent transplacental transfer during pregnancy.

The health implications of prenatal exposure are the subject of a growing and sobering literature. Phthalates have been linked in epidemiological and experimental studies to altered reproductive development, preterm delivery, changes in fetal growth parameters, and disruption of placental function-related genes. Research on the placental methylome and transcriptome suggests that phthalate exposure during the first trimester can reprogram gene regulation at a critical developmental window. Experimental work has also documented oxidative damage and suppression of the Wnt/β-catenin signaling pathway in placental tissue following dibutyl phthalate exposure. For the non-phthalate alternatives, the toxicological record is thinner but increasingly concerning; recent in silico and in vitro studies have flagged diisobutyl adipate, one of the compounds detected in this cohort, for potential lipid metabolism-disrupting effects.

Exposure sources for pregnant women are pervasive and difficult to avoid. Phthalates migrate from PVC food packaging into fatty foods, leach from bottled beverages, and contaminate spices, herbs, and dairy products, as previous surveys of Tunisian and Mediterranean markets by some of the same research groups have documented. Cosmetics and personal care products contribute dermal exposure, with DEP in particular a common ingredient of fragranced formulations. Indoor dust carries plasticizers shed from floorings, wall coverings, and consumer goods, while medical devices can deliver direct doses to vulnerable patients. The Monastir cohort, drawn from middle-class households, underscores that this is not an exposure limited to industrial workers or heavily polluted environments—it is a background exposure affecting the general population of expectant mothers.

The study’s analytical approach deserves attention because measuring parent plasticizers in biological matrices is technically demanding. These compounds are rapidly metabolized in the body into monoester metabolites, which are the usual targets of human biomonitoring in urine. Detecting the parent compounds themselves in placenta, amniotic fluid, blood, and urine, at geometric means ranging from tens to hundreds of micrograms per kilogram, provides a snapshot of circulating chemical burden rather than merely of recent excretion. The multi-matrix design is what allows the exposure continuum to be reconstructed: chemicals present in maternal blood and urine, retained in placental tissue, and dissolved in amniotic fluid trace a pathway from external exposure to the fetal environment.

The authors argue that their findings support further evaluation of both phthalates and their substitutes, and the data give that call force. If DEHT, the flagship replacement, now dominates the exposure profile of pregnant women and penetrates the amniotic cavity, then regulatory success measured by the disappearance of restricted phthalates may mask a regrettable substitution problem, in which one hazardous chemical is swapped for another whose safety profile is unproven in humans. The researchers emphasize the need for improved exposure control, especially for vulnerable populations such as pregnant women and developing fetuses. For now, the study stands as a technical warning: the plastic chemistry of modern life does not stop at the placental barrier, and the burden of proof is shifting to the alternatives that industry hoped would close the chapter on phthalate risk.

Subject of Research: Occurrence and placental transfer of phthalate and non-phthalate plasticizers in maternal and fetal biological matrices

Article Title: Occurrence, distribution and potential sources of phthalate and non-phthalate plasticizers in maternal and fetal-related biological matrices

Article References: Occurrence, distribution and potential sources of phthalate and non-phthalate plasticizers in maternal and fetal-related biological matrices. (n.d.). https://doi.org/10.1007/s10653-026-03540-x

Image Credits: AI Generated

DOI: 10.1007/s10653-026-03540-x

Keywords: phthalates, plasticizers, DEHT, endocrine disruptors, prenatal exposure, placental transfer, amniotic fluid, human biomonitoring, GC-MS/MS, Tunisia, pregnancy, non-phthalate alternatives

The starting point is a technical innovation in how water itself is measured. Rather than relying on irrigation quotas or rainfall records, the researchers quantify effective crop water using actual evapotranspiration, or ETa, derived from the FAO’s WaPOR satellite database. Evapotranspiration captures the water a crop genuinely consumes—the sum of evaporation from soil and transpiration through leaves—making it a far more honest accounting unit than the water that is nominally delivered. Dekadal ETa values are summed to seasonal totals, converted into volumes using intersected administrative and crop-mask areas, and screened with quality flags so that low-confidence satellite pixels are masked rather than imputed. Every governorate-crop-year cell in the resulting panel carries provenance and quality-control documentation, a level of transparency that the authors treat as essential for any policy-facing conclusion.

But measuring water is only half the problem. The other half is causal: how much extra yield does an additional unit of effective crop water actually produce? Simply correlating water use with yields would be misleading, because farmers who irrigate more may differ systematically from those who do not—in soil quality, capital, or institutional access. To solve this, the researchers assembled a governorate-by-crop-by-year panel spanning 2010 to 2024 and identified yield responses using instruments that plausibly shift water availability without directly affecting yields: lagged rainfall anomalies aggregated over upstream basins, with pixels inside the target governorate excluded, and reservoir filling ratios matched to downstream irrigated perimeters through an assignment matrix built from scheme geography and operator records. The logic is elegant. Rain falling upstream of a governorate, and water stored in dams feeding its canals, change how much water crops can consume, but they do not directly change the agronomic or economic conditions inside the governorate itself.

On top of this causal backbone sits the study’s most distinctive component: a fuzzy inference layer that captures the state dependence of water’s marginal value. The idea is grounded in agronomic reality. The same proportional increase in effective crop water does not produce the same yield gain everywhere. Where irrigation delivery is already adequate and technology is efficient, an extra cubic metre buys relatively little; where water stress is high and a binding constraint exists, the same increment can be transformative. To formalise this, the authors built a fuzzy system with three antecedents—water stress, irrigation adequacy, and technology efficiency—each expressed through trapezoidal membership functions that allow a range of values to be unambiguously high or low rather than a single sharp threshold. Three linguistic terms per antecedent generate twenty-seven rules, each encoding an economically and agronomically plausible statement about when marginal returns to water should be larger or smaller. The full rule base, calibrated membership-function parameters, and alternative rule structures are published in the paper’s appendices, an unusually complete disclosure for a component that might otherwise be dismissed as researcher discretion.

The results upend the conventional ranking of where water should go. Average water productivity—the yield per unit of water that dominates most irrigation planning—would direct scarce water toward the regions and crops that already use it most efficiently. But when the state-dependent, causal marginal value of water is estimated instead, the ranking changes. A proportional increase in effective crop water yields different gains across Tunisia’s governorates, and accounting for this heterogeneity reverses some of the priorities implied by average productivity. In practical terms, the water that looks most productive on paper may not be the water that generates the most value at the margin, because productivity averages conceal the local hydrologic and technological states that determine what one more cubic metre is actually worth.

The framework’s second pillar connects agricultural water to human wellbeing. The authors construct governorate-level multidimensional deprivation indicators using household survey and regional accounts data from Tunisia’s National Institute of Statistics, drawing on a fuzzy measurement tradition in poverty analysis that Belhadj has developed in earlier work. They then compare water allocation scenarios against three criteria simultaneously: efficiency, in the total value generated; distribution, in how deprivation changes across territories; and institutional feasibility, scored by a second fuzzy system over five criteria including the metered delivery headroom that existing infrastructure can absorb. This triple test is what elevates the paper from an econometric exercise to a statement about territorial quality of life. Water allocation, the authors argue, should be treated as an environmental-social determinant of wellbeing—one link in a chain that runs from the canal to the household.

The headline finding concerns the trade-off between efficiency and equity. In the scenario accounting, a Rawlsian maximin allocation—one that prioritises the worst-off territories, in the spirit of philosopher John Rawls’s difference principle—retains 90.9 percent of the total value gain that a purely utilitarian allocation would achieve. When the comparison is restricted to crop groups whose elasticity estimates pass over-identification tests, the maximin allocation retains 97.0 percent of the utilitarian gain, while producing a larger reduction in mean deprivation. In other words, the equity-efficiency trade-off that policy makers usually assume to be painful may be remarkably mild: choosing the fairer allocation sacrifices almost nothing in total value while delivering substantially greater reductions in territorial deprivation. For water-stressed Mediterranean governments wrestling with allocation politics, that is a strikingly hopeful result.

The authors are notably careful about what their evidence can and cannot support. The Hansen over-identification test rejects for vegetables and high-value annuals at p equals 0.007, meaning the two instrument families do not identify a common effect for that group; the corresponding coefficient is therefore reported as a diagnostic rather than a validated causal estimate, and the Nabeul citrus value, which uses the same coefficient as a proxy, inherits that limitation. The transmission step from agricultural value added to deprivation is explicitly framed as accounting rather than causal inference: the income-deprivation gradient used to convert value gains into deprivation changes is an estimated cross-sectional association from a single household survey, applied uniformly across governorates because cell sizes do not support separate estimation. The paper claims only the productive and distributional stages of the water-to-wellbeing pathway, not wellbeing outcomes themselves—a restraint that strengthens rather than weakens its credibility.

The robustness apparatus is equally thorough. Sensitivity analyses vary the fuzzy specification, testing triangular and Gaussian membership functions against the trapezoidal baseline and alternative rule structures, and re-estimate the central results without the fuzzy layer entirely, showing that the substantive conclusions survive even when the discretionary component is removed. Leave-one-reservoir-out reconstructions test the storage instrument, alternative SPEI drought horizons test the climate inputs, and cells failing satellite coverage thresholds are masked rather than imputed so that data uncertainty remains traceable and separate from sampling uncertainty. Scenario capacities are checked against one-season metered delivery headroom records held by Tunisia’s regional agricultural development commissariats and agricultural development groups, with each pilot module capped at the minimum of a conservative 200,000 cubic metre ceiling and the actual infrastructure headroom—an upper bound on what existing systems could absorb without new construction.

What emerges is a template that extends well beyond Tunisia. The Mediterranean basin is among the world’s most water-stressed regions, and allocation decisions there are increasingly made under conditions of genuine scarcity, where every cubic metre redirected is a cubic metre taken from someone. By combining satellite-derived water accounting, credible causal identification, transparent state-dependent modelling, and multidimensional deprivation measurement, the framework offers a way to make those decisions with evidence about their distributional consequences rather than averages that hide them. The finding that a fairness-first allocation can retain nearly all of the efficiency gains—while cutting deprivation more deeply—suggests that the hardest choices in Mediterranean water policy may be less agonising than feared, provided the value of water is measured where it actually lands, in the states and places where crops and communities actually live.

Subject of Research: Causal, state-dependent estimation of the marginal value of irrigation water and its distributional effects on territorial quality of life in Tunisia

Article Title: From Crop Water Productivity to Territorial Quality of Life: A Causal, State-Dependent ETa Framework for Water Allocation in Mediterranean Agriculture

Article References: From Crop Water Productivity to Territorial Quality of Life: A Causal, State-Dependent ETa Framework for Water Allocation in Mediterranean Agriculture. (n.d.). https://doi.org/10.1007/s11205-026-03949-y

Image Credits: AI Generated

DOI: 10.1007/s11205-026-03949-y

Keywords: water allocation, evapotranspiration, Tunisia, Mediterranean agriculture, causal inference, instrumental variables, fuzzy logic, multidimensional deprivation, water productivity, quality of life, irrigation policy, satellite data

The research team, led by Fatma Ezzahra Boujmil and colleagues, recruited participants who were undergoing medical examinations to determine their fitness for work, a setting that placed the study squarely at the intersection of psychiatry and occupational medicine. All patients were in a state of clinical euthymia, meaning their mood symptoms were stabilised at the time of assessment. This detail matters scientifically: it demonstrates that even when the florid symptoms of bipolar disorder are controlled, a substantial residue of psychological and functional damage can persist. The average age of participants was 43.8 years, squarely within prime working life, and the gender ratio of 0.53 indicates that men outnumbered women in the sample by nearly two to one, reflecting the composition of the working population examined.

To quantify self-stigma, the researchers used the Internalised Stigma of Mental Illness scale, a validated instrument that probes experiences of alienation, stereotype endorsement, perceived discrimination, social withdrawal and stigma resistance. The average ISMI score in the sample was 2.58, with a standard deviation of 0.87, a level that placed the majority of patients above the threshold conventionally used to identify meaningful internalised stigma. Functioning, meanwhile, was assessed with the Functioning Assessment Short Test, a tool designed to capture the real-world capacities that matter most to patients: managing a job, maintaining relationships, handling finances, and sustaining the cognitive abilities needed for independent living. The results were sobering. Overall functional impairment was present in 89.6 percent of patients, with occupational functioning affected in 64.2 percent and cognitive functioning in 50 percent, making these two domains the most heavily damaged.

The demographic and clinical correlates of self-stigma painted a picture of vulnerability concentrated among specific groups. Average ISMI scores were significantly higher in patients under the age of 40, in women, in those who were single or divorced, and in patients with low incomes. Each of these factors plausibly amplifies the psychological weight of a psychiatric diagnosis. Younger patients may still be constructing their professional and personal identities when the illness strikes, making the internalised label more corrosive. Women in the Tunisian working context may face layered expectations that intensify the sense of falling short. Social isolation, whether through singleness or divorce, removes protective relationships that might otherwise buffer negative self-beliefs, and financial hardship compounds the sense of diminished worth and limited horizons.

Occupational characteristics also left their mark. Patients working atypical hours, such as night shifts or irregular schedules, carried significantly higher self-stigma scores than those with standard working patterns. This finding resonates with a growing body of occupational health research showing that disrupted circadian rhythms and unstable schedules are particularly punishing for people with mood disorders, since sleep regularity is a cornerstone of bipolar disorder management. An atypical schedule may also signal to patients themselves that they occupy a marginal position in the world of work, feeding the internal narrative of difference and deficiency. Early onset of bipolar disorder was another correlate of heightened self-stigma, consistent with the idea that the longer a person lives with the illness, and the earlier it interrupts the formation of adult identity, the deeper the internalised stereotypes can root.

One of the most clinically actionable findings concerned treatment. Patients who had not received cognitive behavioural therapy showed significantly higher self-stigma scores than those who had. CBT is specifically designed to identify and challenge distorted beliefs about the self, and its protective association here suggests that structured psychological therapy may do more than alleviate mood symptoms; it may actively dismantle the internalised prejudice that patients absorb from their social environment. In health systems where psychiatric care is often reduced to medication management and brief consultations, the study offers an argument for embedding psychotherapy into routine care for bipolar disorder, not as a luxury but as a targeted intervention against a measurable driver of disability.

The statistical heart of the paper lies in the correlation between self-stigma and functioning. Higher ISMI scores were significantly associated with greater overall functional impairment on the FAST, with the strongest links involving the perceived discrimination and social withdrawal subscales. This pattern is theoretically coherent: a patient who expects rejection and consequently withdraws from colleagues, friends and family loses the opportunities to test those expectations against reality, and the withdrawal itself erodes social skills, professional networks and confidence. Functioning deteriorates not because the illness has flared, but because the patient has pre-emptively shrunk their own world. The authors emphasise that this dynamic is widely underestimated, particularly in the workplace, where the quiet retreat of a capable employee may be misread as laziness or lack of ambition rather than recognised as the downstream effect of internalised stigma.

Methodologically, the study has the strengths and limits inherent to its cross-sectional design. Because all data were collected at a single point in time, the direction of causality between self-stigma and impaired functioning cannot be definitively established; it is plausible that each feeds the other in a vicious cycle, with functional difficulties confirming negative self-beliefs and those beliefs in turn undermining functional recovery. The sample of 106 patients recruited through fitness-for-work examinations is also a selected population, since severely disabled patients may no longer be employed at all, potentially leading the study to underestimate the true burden of self-stigma in the broader bipolar population. Nevertheless, the consistency of the associations, the use of validated instruments, and the ethical rigour of the protocol, which adhered to the Declaration of Helsinki and was approved by the ethics committee of Charles Nicole Hospital, lend the findings considerable weight.

The implications reach well beyond Tunisia. Bipolar disorder affects roughly one percent of the population worldwide and primarily strikes people of working age, meaning that its economic and social costs are concentrated in the very decades when individuals are building careers and families. If more than half of working patients carry internalised stigma and nearly nine in ten show functional impairment even in remission, then treating mood symptoms alone addresses only part of the problem. The authors conclude that the high prevalence of self-stigma and its association with impaired functioning highlight the need to address contributing factors, particularly in the workplace, in order to maintain employment and limit the adverse impact on the progression of the disorder. That agenda could include workplace mental health education to dismantle public stereotypes before they are internalised, schedule accommodations for patients with mood disorders, routine screening for internalised stigma in psychiatric follow-up, and wider access to cognitive behavioural therapy. What this study makes clear is that the battle against bipolar disorder is fought on two fronts, and the quieter one, inside the patient’s own mind, may be the one that decides whether recovery translates into a functioning life.

Subject of Research: Self-stigma and functional impairment in working patients with bipolar disorder

Article Title: Bipolar disorder, self-stigma and functioning: a cross-sectional study conducted among a working population in Tunis

Article References: Bipolar disorder, self-stigma and functioning: a cross-sectional study conducted among a working population in Tunis. (n.d.). https://doi.org/10.1186/s12888-026-08727-1

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08727-1

Keywords: bipolar disorder, self-stigma, internalised stigma, functioning, occupational health, workplace mental health, cognitive behavioural therapy, Tunisia, cross-sectional study, psychiatry, social withdrawal, perceived discrimination

Treated wastewater, often abbreviated TWW, is municipal or industrial effluent that has passed through purification systems and been deemed suitable for reuse. In water-scarce countries such as Tunisia, it represents a lifeline for agriculture, supplementing or replacing scarce freshwater for fodder and cereal crops. Yet the treatment process is not perfect. Residual pollutants, dissolved salts, heavy metals in trace amounts, and organic chemicals can survive conventional treatment and find their way into soil, plant tissues, and ultimately seeds. The concern is cumulative: a single season of TWW irrigation might be harmless, but what about two, three, or more? If contaminants build up in the edible parts of the plant, the nutritional and biochemical integrity of the harvest could erode over time, quietly and invisibly.

That cumulative question is precisely what the Tunisian team set out to address. Rather than examining a single growing season, the researchers tracked barley plants irrigated with treated wastewater for one, two, and three consecutive cycles, and then analysed the seeds those plants produced. Barley is an ideal candidate for such an investigation. It is a staple cereal in many semi-arid regions, is commonly grown under deficit irrigation, and its seeds are both an animal feed and, in various forms, a human food ingredient. Any degradation in seed quality would therefore ripple directly through local food and feed chains.

The analytical focus of the study fell on two broad classes of compounds. The first is primary metabolites, the fundamental molecules of plant life, including sugars, proteins, and other compounds that determine the basic nutritional value of a seed. The second is the secondary metabolite profile, dominated in barley by phenolic compounds, a chemically diverse family of molecules that plants synthesize as part of their defence machinery against stress, pathogens, and oxidative damage. Phenolics also matter to human health, since many act as antioxidants, and they influence the flavour, colour, and processing characteristics of cereal products. Measuring both classes across successive cycles allowed the researchers to ask whether TWW irrigation erodes nutrition, alters defence chemistry, or both.

The headline result is strikingly positive. Across one, two, and three cycles of treated wastewater irrigation, seed production increased, meaning the plants did not merely survive the recycled water but appeared to benefit from it, likely because the treated effluent carries nutrients that freshwater irrigation does not. More importantly for food quality, the primary metabolites of the harvested seeds did not deteriorate. The basic nutritional composition of the barley grain remained intact regardless of how many seasons of TWW irrigation the parent plants had experienced. For a region where treated wastewater is increasingly the only reliable irrigation source, this is a consequential finding.

The secondary metabolite picture was more nuanced, and arguably more interesting. The researchers found that TWW irrigation induced a cycle-dependent phenolic metabolism, meaning the pattern of phenolic compounds in the seeds changed in ways that tracked the number of irrigation cycles. Rather than degrading the bioactive profile of the grain, prolonged exposure to treated wastewater appeared to promote the defence machinery embedded in the produced seeds. This is consistent with a well-established principle in plant physiology: mild chemical or environmental stress can trigger the synthesis of protective secondary metabolites. The residual solutes and compounds in treated wastewater, present at low levels, may act as a subtle elicitor, prompting the developing seeds to invest more heavily in phenolic defences.

Crucially, the team concluded that there were no negative cumulative effects on either the primary or the secondary metabolites of the seeds. In plain terms, three seasons of recycled water did not poison the grain, deplete its nutrition, or destabilize its biochemistry. The metabolic changes that did occur were directional and, as far as the measured parameters showed, benign or even advantageous. The authors summarize the practical implication directly: long-term irrigation with treated wastewater does not affect the nutritional quality of barley seeds. For policymakers weighing the expansion of wastewater reuse programs, that sentence may carry more weight than any laboratory measurement.

Why does this matter on a global scale? Water scarcity is no longer a regional problem confined to the Sahel or the Middle East. According to widely cited assessments, a growing share of the world’s population lives under conditions of chronic water stress, and agriculture consumes the majority of freshwater withdrawals in most arid countries. Treated wastewater is one of the few genuinely expandable water resources available, because cities produce more of it every year as populations grow. Countries such as Israel, Spain, Australia, and several Gulf states already reuse large volumes of treated effluent for agriculture, and many others are drafting regulations to follow. The chief scientific objection has always been uncertainty about long-term, multiseason effects on crops and soils. Studies like this one, which explicitly test cumulative exposure rather than a single season, chip away at that uncertainty with empirical data.

The study also carries a broader methodological lesson. Much of the existing literature on wastewater irrigation focuses on what accumulates in soils, roots, or leaves, or on crop yields alone. Seed biochemistry is a quieter, more demanding measurement, but it is arguably the most relevant endpoint for food safety, because seeds are the harvested commodity and the vehicle through which any accumulated contaminants would reach consumers or livestock. By focusing on the produced seeds and interrogating both primary and secondary metabolism, the Tunisian team targeted the question that actually matters to farmers, millers, and regulators. The work was carried out at the Dry Land and Oases Cropping Laboratory of the Arid Regions Institute of Medenine, with support from the Tunisian Ministry of Higher Education and Scientific Research, a reminder that some of the most directly relevant water research is produced in the very regions where the stakes are highest.

None of this means treated wastewater is a risk-free elixir. The composition of treated effluent varies with the source sewage, the treatment technology, and the local regulations governing what may be discharged, and results from barley in a Tunisian context cannot be transplanted uncritically to other crops, soils, or treatment standards. Contaminants that were not measured, or effects that emerge only after more than three cycles, remain open questions, and the authors themselves frame their conclusions within the scope of the metabolites they examined. But the study delivers something genuinely valuable: evidence that, at least for a hardy cereal under realistic multiyear conditions, recycled municipal water can raise yields, preserve nutritional quality, and even nudge the plant’s chemical defences in a favourable direction. In a century defined by water scarcity, that combination of productivity and safety may prove to be one of agriculture’s most important bargains.

Subject of Research: Cumulative biochemical effects of treated wastewater irrigation on barley seed metabolites

Article Title: Biochemical responses of barley seeds to the cumulative effects of treated wastewater irrigation

Article References: Ghanmi, A., Ben Ali, S., Boussora, F., Triki, T., Assissila, S., Bouhamed, T., & Guasmi, F. (2026). Biochemical responses of barley seeds to the cumulative effects of treated wastewater irrigation. BMC Plant Biology. https://doi.org/10.1186/s12870-026-10060-4

Image Credits: AI Generated

DOI: 10.1186/s12870-026-10060-4

Keywords: treated wastewater, barley, seed quality, phenolic compounds, primary metabolites, secondary metabolites, water reuse, agriculture, water scarcity, plant biochemistry, Tunisia, BMC Plant Biology

The experiment took place at the Lafareg research farm in Beja, a sub-humid zone receiving around 558 millimeters of rain annually. Twenty-eight five-month-old fat-tailed Barbarine ram lambs, weighing on average 22.5 kilograms, were divided into four groups of seven, balanced by body weight. Two groups were housed in stalls and fed 400 grams of concentrate per day with oat hay available freely. The other two groups rotationally grazed a cultivated vetch pasture, receiving the same daily concentrate allowance. Within each system, one group received an extra 100 grams of acacia foliage per head per day, offered an hour before the main feed, while the other received none. The acacia is a tannin-rich legume widespread in Tunisia, containing roughly 60 grams of condensed tannins per kilogram of dry matter, and it was chosen as a natural alternative to the antibiotic growth promoters that have been banned in many regions over concerns about toxic residues in meat.

The chemical analysis of the feeds explains much of what followed. The fresh vetch grass contained 21.9 percent crude protein, more than three times the 6.7 percent found in the hay offered indoors. The concentrate supplied 16.2 percent protein, and the acacia itself 12.6 percent. Because grazing lambs could harvest this protein-dense forage directly, their total dry matter intake reached 1,342 to 1,442 grams per day, compared with only 839 to 960 grams for the stall-fed animals. Grazing lambs derived about 70 percent of their intake from forage, eating nearly a kilogram of vetch daily, while stall-fed lambs drew 47 percent of their ration from concentrate and consumed less than half as much forage as their pastured counterparts.

Growth told the same story in striking numbers. Lambs on pasture gained an average of 180 grams per day, while stall-fed lambs gained 141 grams. The extremes were even more revealing: the fastest-growing group was pasture-fed lambs without acacia, at 192 grams per day, and the slowest was stall-fed lambs without acacia, at 110 grams. The statistical interaction between feeding system and tannin supply was significant, meaning the effect of the supplement could not be judged in isolation from the basal diet. Indoors, where nitrogen arrived as true protein in hay and concentrate, the tannins appeared to offer a slight benefit, consistent with the well-documented ability of condensed tannins at low doses to bind dietary protein in the rumen and protect it from degradation, thereby increasing the flow of amino acids to the small intestine.

On pasture, however, the chemistry worked differently. The nitrogen in fresh vetch is largely in soluble form, which tannins cannot bind in the same protective way, and the vetch biomass itself already carries measurable tannins, averaging around 360 milligrams of catechin equivalents per kilogram of dry matter. Adding acacia on top of this tannin-bearing forage apparently pushed total tannin intake toward a range where the compounds begin to work against the animal, suppressing feed intake, rumen microbial activity, and digestive enzyme function. The authors note that previous work has shown low to moderate tannin inclusion can be neutral or beneficial, while larger amounts reduce nutrient digestibility and growth. The lesson is that tannins are not a simple additive but a dietary variable whose consequences hinge on context.

At slaughter, the pasture advantage carried through to the carcass. Grazing lambs were heavier at slaughter, had higher empty body weights, and produced both heavier hot and cold carcasses. Their commercial dressing percentage averaged 45.8 percent against 40.5 percent for stall-fed lambs. The non-carcass components also reflected the different lives the animals had led: grazing lambs developed larger red organs, including heart, liver, and lungs, and heavier kidneys, a pattern the authors attribute to the greater physical activity of animals ranging over pasture. Testis weight was higher in both pasture groups and in the tannin-supplemented stall group compared with unsupplemented stall-fed lambs, while gut weights were similar across all treatments.

Within the carcass itself, the feeding system shaped the proportions of most commercial joints, including the leg, shoulder, loin, breast, and, notably, the signature fat tail of the Barbarine breed. Pastured lambs carried tails averaging 950 grams, about 6.05 percent of carcass weight, compared with 615 grams and 4.7 percent in stalled lambs. Muscle proportions were highest in unsupplemented stall-fed lambs, while fat deposition was greater in both pasture groups and in tannin-supplemented stall lambs, an outcome that runs counter to earlier studies in which grazing animals, burning more energy on movement, laid down less fat. Across all groups, the leg and shoulder together accounted for roughly 34 percent and 20 percent of the tailed carcass, supporting the principle of anatomic harmony, the idea that joint proportions remain consistent within a breed regardless of rearing conditions.

Perhaps the most commercially reassuring finding concerned the meat itself. Neither the feeding system nor the acacia supplement produced meaningful differences in pH measured one hour and 24 hours after slaughter, in water loss during cooking, in color coordinates of lightness, redness, and yellowness, or in the chemical composition of the longissimus dorsi muscle, including dry matter, ash, protein, and intramuscular fat. Lightness values averaged 43.5 across all groups, comfortably above the threshold of 34 at or above which meat is deemed acceptable by 95 percent of consumers. Trained panelists who roasted samples from each group and scored tenderness, juiciness, flavor, and overall acceptability on nine-point scales found the meat moderately tender, between 5.7 and 6, and moderately juicy, between 5.1 and 5.8, with no distinctive off-flavors and no differences among treatments. Every sample was judged acceptable.

The broader significance of the study lies in its demonstration that natural feed additives cannot be evaluated in a vacuum. Condensed tannins have attracted intense interest as antioxidants, antimicrobials, and rumen-protective compounds that could replace banned growth promoters, but their biological activity is a double-edged sword, beneficial at low doses and detrimental at high ones. This trial shows that the dose that matters is the total dose delivered by the whole diet, not just the supplement spooned out on top. A hundred grams of acacia that gently shields protein in a hay-based ration becomes an excess burden when the lambs are already browsing tannin-bearing vetch. For producers of fat-tailed breeds like the Barbarine, the practical takeaways are twofold: well-managed grazing on protein-rich cultivated pasture remains the most effective route to faster growth and heavier carcasses, and tannin supplementation should be matched carefully to the basal forage. The authors call for further research to define optimal inclusion levels and dose-response relationships under different feeding conditions, work that will help determine whether tannin-rich foliage can take its place as a reliable tool in sustainable, antibiotic-free lamb production.

Subject of Research: Effects of Acacia cyanophylla condensed tannin supplementation and grazing versus feedlot feeding on growth, carcass traits, and meat quality of Barbarine lambs

Article Title: Effects of Acacia cyanophylla-condensed tannin supply on growth performance, carcass characteristics and meat quality of fat-tailed Barbarine lambs studied under grazing conditions or in a feedlot

Article References: Yagoubi, Y., Atti, N., Smeti, S., Abidi, J., & Mahouachi, M. (2026). Effects of Acacia cyanophylla -condensed tannin supply on growth performance, carcass characteristics and meat quality of fat-tailed Barbarine lambs studied under grazing conditions or in a feedlot. Archives Animal Breeding, 69(3), 495-505. https://doi.org/10.5194/aab-69-495-2026

Image Credits: AI Generated

DOI: 10.5194/aab-69-495-2026

Keywords: Barbarine lambs, Acacia cyanophylla, condensed tannins, grazing, feedlot, carcass quality, meat quality, sheep nutrition, sustainable livestock, vetch pasture, animal growth, Tunisia

The timing of the study is significant. Tunisia launched its national e-learning Mentor initiative, known as ELM, as a major government investment in AI-based educational resources for the 2024 to 2025 academic year. The researchers, Ichrak Riahi and Wafa Battikh of the University of Jendouba and Hanen Khanchel of the University of Carthage, collected their survey data between March and May 2024, capturing student perceptions at a pivotal moment just before full implementation. This gave them a snapshot of expectations and attitudes at the exact point when the country’s universities were committing to a sweeping digital transformation.

The theoretical backbone of the study is the Technology Acceptance Model, or TAM, first proposed by Fred Davis in 1989 and still one of the most widely used frameworks for predicting how people adopt new technologies. TAM centers on two constructs: perceived usefulness, meaning the degree to which a user believes a technology will improve their performance, and perceived ease of use, meaning how effortless the technology feels to operate. The Tunisian team extended this classic model by adding electronic trust, satisfaction, and student engagement as downstream outcomes, and then introduced a variable rarely tested in this context: the ethical perceptions students hold about smart learning tools.

The statistical results reveal a clear hierarchy of influence. Perceived usefulness significantly boosted electronic trust, with a path coefficient of 0.122, while perceived ease of use had an even stronger effect at 0.257. In plain terms, students who found smart technologies helpful and simple to navigate were markedly more inclined to trust them. Ease of use appears to work by making the learning experience feel manageable and intuitive, reducing the sense of risk that often accompanies unfamiliar digital systems. This finding aligns with a broad international literature showing that usability is a gateway to confidence in online environments.

Trust, in turn, proved to be the single most powerful driver of engagement in the entire model. The path from electronic trust to student engagement registered a coefficient of 0.352, and trust also strongly predicted satisfaction at 0.416. Satisfaction itself fed back into engagement with a coefficient of 0.383, suggesting a virtuous cycle: students who trust a platform become satisfied with it, and satisfied students invest more of themselves in the learning it delivers. The researchers describe satisfaction as playing a complementary mediating role, channeling some of trust’s influence on engagement while trust also exerts a direct effect of its own.

The most novel contribution of the study lies in its treatment of ethics. The researchers hypothesized that students’ ethical perceptions, covering concerns such as data privacy, algorithmic bias, and the responsible use of artificial intelligence, would moderate several of these relationships. The results were nuanced. Ethical perception did not significantly moderate the link between trust and engagement, with a p-value of 0.111 falling short of the 0.05 threshold. However, it did significantly strengthen the effect of trust on satisfaction, with an unstandardized coefficient of 0.0918 and a p-value of 0.045, and it magnified the effect of satisfaction on engagement, with a coefficient of 0.1091 and a p-value of 0.021.

These moderation findings carry a striking implication: ethics is not a peripheral concern but an amplifier of the emotional and motivational pathways that make students stick with smart learning technology. When students believe a platform respects their data and operates fairly, their trust converts more readily into satisfaction, and their satisfaction converts more readily into deep engagement. The researchers draw on fairness heuristic theory to explain this pattern, noting that people gravitate toward entities they perceive as ethical and that ethical behavior signals respect for user rights and values. When a technology’s conduct aligns with a student’s moral identity, attitudes toward that technology warm considerably.

The study is candid about its limits. The cross-sectional, self-report design cannot establish causality, and the convenience sample, while deliberately diverse across disciplines and degree levels, was drawn from a single country during a single semester. The moderating effects, though statistically significant, explained only modest amounts of additional variance, and the overall model accounted for a small portion of engagement overall, hinting that other unmeasured factors, from instructor quality to home internet access, also shape how students engage with smart tools. The authors call for longitudinal designs, cross-cultural comparisons, and qualitative work using interviews and focus groups to probe how students actually reason about the ethics of educational AI.

For policymakers and university leaders, the practical recommendations are concrete. The authors urge the creation of national ethical standards for educational AI, addressing algorithmic transparency, data privacy, and responsible use, developed by working groups that include educators, technologists, and ethicists. They recommend funding digital literacy and ethics programs embedded in curricula and professional development, and they stress equity of access, arguing that digital inequalities across Tunisian universities must be addressed so that students with disabilities and those at under-resourced institutions are not left behind. Educators, meanwhile, are encouraged to position AI as a support for teaching rather than a substitute for it, keeping human feedback loops at the center of the learning experience.

What makes this research resonate far beyond Tunisia is its reminder that the adoption curve of educational technology is not purely a matter of features and functionality. As universities worldwide pour resources into AI tutors, virtual reality laboratories, and adaptive learning platforms, the Tunisian findings suggest that the return on those investments will hinge on trust built through usability, satisfaction earned through genuine usefulness, and, crucially, ethical credibility that students can perceive and verify. In an era when students increasingly weigh the moral dimensions of the AI tools they use, institutions that treat ethics as a design requirement rather than an afterthought may find their smart learning technologies not just adopted, but embraced.

Subject of Research: The impact of smart learning technologies on student engagement, e-trust, and satisfaction in higher education, with ethical perceptions as a moderator

Article Title: Exploring the impact of smart learning technologies on student engagement, e-trust, and satisfaction in higher education institutions: the moderating role of ethical perceptions

Article References: Riahi, I., Battikh, W., & Khanchel, H. (2025). Exploring the impact of smart learning technologies on student engagement, e-trust, and satisfaction in higher education institutions: the moderating role of ethical perceptions. Journal of New Approaches in Educational Research, 14(1), Article 21. https://doi.org/10.1007/s44322-025-00039-8

Image Credits: AI Generated

DOI: 10.1007/s44322-025-00039-8

Keywords: smart learning technology, higher education, student engagement, e-trust, technology acceptance model, ethical perception, e-learning, Tunisia, artificial intelligence in education, structural equation modeling, student satisfaction, data privacy

The study, led by Zaineb Hamzaoui and colleagues at Charles Nicolle Hospital and the Faculty of Medicine of Tunis, drew on a national sentinel network spanning all 24 Tunisian governorates. The network combined 85 community-based influenza-like illness sites in primary health care centers with 11 hospital-based severe acute respiratory infection departments at university hospitals, all operating under WHO 2014 case definitions. Between November 1, 2021 and September 30, 2025, the laboratory received 8,635 nasopharyngeal swabs through this system and confirmed 1,020 SARS-CoV-2 infections by real-time RT-PCR, an overall positivity rate of 11.8 percent. Under the study protocol, every positive sample with a cycle threshold value below 30 was attempted for whole-genome sequencing, a threshold chosen because higher Ct values typically yield incomplete genomes and unreliable lineage assignment.

The technical workflow relied on viral RNA extraction with the Chemagic 360 instrument, library preparation using Illumina RNA Prep with Enrichment and the Respiratory Virus Oligo Panel, and sequencing on an Illumina iSeq 100. Consensus genomes were built against the Wuhan-Hu-1 reference, quality-checked in Nextclade, and cross-validated with Genome Detective before Pango lineage and GISAID clade assignment. Of the 820 sequencing attempts, 669 genomes—81.6 percent—passed quality control and were deposited in GISAID, representing 65.6 percent of all RT-PCR-confirmed infections diagnosed at the reference laboratory during the study period. That proportion is unusually high for a middle-income setting and approaches international benchmarks for adequate genomic surveillance. Mapped read proportions were consistently high, typically at or above 0.9, lending confidence to the mutation counts derived from the data.

Geographic representativeness emerged as one of the study’s strongest findings. The researchers computed an equity index for each governorate, comparing its share of national sequences with its share of national positives, and found that the distribution of sequences closely tracked the distribution of cases, with a Pearson correlation of 0.992 across the 24 governorates. Median governorate-level sequencing coverage among RT-PCR-positive samples was 60.0 percent, with an interquartile range of 28.6 to 88.9 percent. Coverage was generally higher in densely populated coastal regions, often reaching 80 percent or more, while several central and southern areas fell below 40 percent and a few interior governorates contributed no genomes at all. The equity index flagged relative under-sampling in 12 of 24 governorates and over-sampling in one, though most imbalances occurred in governorates with very few confirmed cases, limiting their impact on the national picture.

Turnaround time—the interval between specimen collection and genome submission to GISAID—was more heterogeneous. The overall median was 99 days, with an interquartile range of 60 to 146 days, and only 43.5 percent of genomes were submitted within 90 days of collection. Nearly 15 percent of sequences experienced delays exceeding 200 days, concentrated among genomes collected in 2022 and 2023. The picture improved markedly in later years: median turnaround fell to 38 days in 2024 and 67 days in 2025, and no delays beyond 200 days were observed among genomes collected in 2024 or 2025. Within higher-volume governorates, median turnaround ranged from 22 days in Sousse to 167 days in Ariana. The authors note that international guidance recommends sharing sequence data within roughly 21 days of collection, a benchmark that accelerates variant detection and risk assessment, and they point to streamlining specimen transport and bioinformatics workflows as routes to improvement.

The lineage data reveal a strikingly dynamic viral landscape. XBB and its descendants accounted for the largest share of genomes at 25.0 percent, followed by BA.4/BA.5 lineages at 22.1 percent, JN.1 and its descendants at 13.6 percent, BA.2 at 12.6 percent, and BA.1 at 9.0 percent. Residual Delta lineages represented 8.7 percent, with the remaining 9.1 percent belonging to other lineages, including pre-Omicron variants and non-XBB recombinants. Nearly all genomes fell within the Omicron-associated GISAID clades GK and GRA. Within these broad categories, a handful of individual Pango lineages—most notably BA.2, BA.5.2, AY.122, BA.1.1, BA.5.2.20, and XBB.2.3.11—contributed the largest shares, while most Omicron sublineages each accounted for less than 2 percent of genomes, a long tail of rare variants underlying the aggregated groups.

The mutational trajectories documented in the dataset chart the virus’s continuing diversification on the Omicron backbone. Early Delta, BA.1, and BA.2 genomes carried roughly 30 to 50 amino acid substitutions relative to the Wuhan-Hu-1 reference, whereas later XBB and JN.1 sublineages frequently exceeded 70 to 80 substitutions, with some recombinants approaching 100. Amino acid deletions also accumulated progressively, reaching 18 to 19 in some JN.1-related genomes. Phylogenetic placement using the Nextclade reference tree showed that Tunisian genomes did not form country-specific outlier clusters but were fully embedded within globally circulating clades: early sequences from late 2021 clustered in Delta clade 21J, while subsequent genomes distributed across successive Omicron clades from 21K and 21L through 22B and 22E to the 23D/23E and 24A through 24H clades corresponding to XBB and JN.1 lineages. This pattern, consistent with repeated introductions rather than long-term local persistence of restricted clusters, mirrors findings from other North African and Mediterranean settings.

Epidemiological curves tied each wave of positivity to the establishment or replacement of a dominant lineage. Testing volumes rose recurrently each winter, peaking in January 2023, but SARS-CoV-2 positivity followed its own rhythm, with substantial peaks in mid-2022 and mid-2023 alongside winter surges. The BA.5.2 wave crested in May through July 2022, the recombinant XBB.2.3.11 drove a surge in July and August 2023, and a JN.1 wave in winter 2023-2024 gave way to extensive circulation of JN.1 offspring through 2024 and 2025. The authors emphasize that these off-season outbreaks underscore that SARS-CoV-2 has not yet settled into the strict winter seasonality of influenza and can still generate waves whenever new immune-evasive variants emerge. The dynamics broadly paralleled those observed across North Africa, the Eastern Mediterranean, and Southern Europe, where KP.3.1.1 and related JN.1 descendants predominated by mid-2024.

Because the sequencing platform sat within a multiplex respiratory surveillance framework, the study also captured viral co-detections. Among the 669 sequenced SARS-CoV-2-positive samples, 92—13.8 percent—showed co-detection with at least one other respiratory virus. Rhinovirus was the most frequent companion at 2.8 percent, followed by adenovirus at 1.5 percent, influenza A/H3N2 at 1.3 percent, and respiratory syncytial virus at 1.2 percent. Influenza viruses were involved in 16 co-detections, or 2.4 percent of sequenced cases. The predominance of rhinovirus aligns with pandemic-era literature showing that rhinovirus and enterovirus persisted or re-emerged earlier than enveloped respiratory viruses, which were markedly suppressed during the first phase of the COVID-19 pandemic. Influenza detections, absent during the 2020-2021 season, resumed once broader multiplex testing was reintroduced from the 2021-2022 season onward.

The authors are candid about the limitations. Clinical metadata were available for only 46.5 percent of sequenced cases, and age data were missing for 8.8 percent of patients, with completeness varying by year and governorate. The Ct-based inclusion criterion may have biased the cohort toward higher viral load infections and earlier sampling timepoints, and the phylogenetic component relied on reference-tree placement rather than de novo reconstruction, precluding formal inference of transmission clusters or introduction counts. The dataset also reflects the ILI/SARI sentinel sampling frame rather than the entirety of national COVID-19 testing. Even so, the study demonstrates that an influenza-based sentinel network in a middle-income country can deliver sustained, geographically near-proportional SARS-CoV-2 genomic surveillance at high sequencing quality. By combining near-proportional sampling across governorates with multipathogen testing, the Tunisian platform offers a scalable model for tracking future SARS-CoV-2 variants and other emerging respiratory pathogens, positioning integrated sentinel surveillance as a pragmatic foundation for pandemic preparedness in the region.

Subject of Research: Genomic surveillance of SARS-CoV-2 Omicron lineage dynamics in Tunisia

Article Title: Genomic surveillance of SARS-CoV-2 in Tunisia during the omicron era: insights from the national influenza & other respiratory viruses laboratory

Article References: Hamzaoui, Z., Ferjani, S., Bouchouicha, T., Charaa, L., Landolsi, I., Medini, I., Ben Ali, R., Chammam, S., Abid, S., Kanzari, L., Ben Sassi, M., Trabelsi, S., & Boutiba Ben Boubaker, I. (2026). Genomic surveillance of SARS-CoV-2 in Tunisia during the omicron era: insights from the national influenza & other respiratory viruses laboratory. Virology Journal, 23(1), Article 215. https://doi.org/10.1186/s12985-026-03262-7

Image Credits: AI Generated

DOI: 10.1186/s12985-026-03262-7

Keywords: SARS-CoV-2, Omicron, genomic surveillance, Tunisia, whole-genome sequencing, Pango lineages, GISAID, JN.1, XBB, respiratory viruses, influenza sentinel network, co-detection

Thuburbo Majus lies roughly sixty kilometers southwest of Tunis, near the modern town of El Fahs, in a fertile agricultural zone that helped the city flourish from the first through the seventh centuries CE. Granted civic status probably under Augustus and later elevated to the rank of honorary colony under Commodus in 188 CE, the city accumulated temples, a forum, a Capitolium and public baths. The Winter Baths, covering approximately 1,600 square meters in the urban core, grew through an asymmetrical layout born of repeated additions and renovations. An inscription studied in earlier scholarship allowed archaeologists to attribute the original construction to the first half of the third century CE, with major restoration phases in the late fourth to early fifth centuries and again in the early sixth century, before the building’s eventual abandonment.

The research team, led by Aida Zaddem and Karima Zoghlami with colleagues including Rafael Fort of the Institute of Geosciences in Madrid, sampled nine distinct stone lithotypes and thirty-two mortar specimens from walls across the complex. Petrographic thin sections, stained to distinguish calcite from dolomite, revealed the full petrographic identity of each rock: a nummulite-rich Oligocene packstone, travertine, ferriferous recrystallized micrite, a dense Jurassic limestone marketed even today as Aziza-type marble, a porous bioclastic calcarenite, speleothem calcite, an Eocene biomicrite, a Jurassic pelsparite and a glauconitic grainstone. Each of these was matched to a specific architectural role, and the match turned out to be strikingly deliberate.

The analysis exposed a procurement strategy of remarkable economy and precision. Nummulitic limestone and travertine, quarried within a radius of less than twenty kilometers, formed the monumental framework, minimizing transport costs while providing the mechanical strength the main structures required. The heating system told an even sharper story. The vault of the main furnace, the praefurnium, was built exclusively of fired bricks bonded with cocciopesto mortar, a technical imperative given that flames there could reach between 800 and 1,000 degrees Celsius, hot enough to decarbonate and disintegrate any carbonate stone. Yet the support pillars and the cheeks of the flue channels, made of glauconitic limestone and ferriferous micrite, remained intact, demonstrating that builders understood empirically that temperatures at those positions stayed below the limestone destruction threshold of roughly 600 to 800 degrees.

Specialization extended to the prestigious elements. Highly porous calcarenite, with porosity reaching 25 to 27 percent, was reserved for arches and voussoirs where lightness mattered, while marbles were chosen for columns, where hardness and visual prestige dominated. The mortars, meanwhile, fell into two great families: lime-based and gypsum-based. The lime mortars split into three subgroups that map directly onto Roman hydraulic science. The first type, moderately hydraulic, mixed a lime binder with siliceous and carbonate aggregates likely drawn from the sands of the nearby Oued Miliane and Oued Medjerda rivers; their hydraulicity came from clay impurities in marly limestone calcined into the binder itself, forming calcium silicate and aluminate hydrates without any added ceramics.

The second type was entirely different in purpose: non-hydraulic air-lime coatings built in multiple layers, their aggregates composed of crushed marble and vein calcite, with grain size progressively decreasing from the wall inward toward the finished surface. These white to beige plasters carried the highest calcium oxide contents of the entire assemblage, averaging around 57 percent, and their chemistry shows almost no silica, alumina or iron, confirming a purely aerial binder. Fascinatingly, one intermediate layer in some samples shows a faint reddish hue and weak hydraulic character, hinting that builders may have sprinkled finely ground ceramic powder, an artificial pozzolana, into that single stratum to fine-tune its properties. This kind of marble-dust coating echoes the plaster recipes that Vitruvius described for polished wall finishes.

The third lime type is the showpiece: highly hydraulic cocciopesto, loaded with ceramic fragments up to a centimeter across, used to waterproof the frigidarium pool and a decantation basin. Thermal analysis sealed the classification. The ratio of carbon dioxide to water lost during heating, a standard proxy for hydraulicity, dropped below 5 percent for these mortars, the signature of strongly hydraulic systems in which pozzolanic reactions between lime and fired clay generate abundant calcium silicate hydrates. Trace elements told the same story from another angle: cocciopesto samples carried the highest concentrations of niobium, rubidium, zirconium, zinc, chromium and vanadium, the geochemical fingerprint of silicate-rich ceramic aggregate. A fourth family of gypsum-based mortars, dominated by sulfate phases with sulfur trioxide averaging over 63 percent, came from the abandonment phase, when builders opportunistically exploited abundant local Triassic gypsum deposits even in structural roles that gypsum was traditionally never asked to play.

Perhaps the most consequential findings are chronological. Mortar fingerprinting forced a re-evaluation of the building’s accepted construction sequence. A joint mortar from the latrines, previously assigned to the first renovation phase of the late fourth century, proved compositionally indistinguishable from the original third-century mortars, raising the possibility that the latrines were part of the initial design and only internally modified later. A coating from the corridor wall, attributed to the original phase on architectural grounds, instead matches the crushed-marble plasters of the late-fourth-century renovation. Most strikingly, the bottom of the decantation basin, built with original-phase cocciopesto, contrasts with its upper walls of gypsum-bonded rubble, suggesting the basin’s base belongs to the first construction campaign while its upper portion was added during abandonment, a functional repurposing of the water system. The authors caution that these reinterpretations remain preliminary until confirmed by absolute dating methods such as optically stimulated luminescence or isotopic analysis.

The study also carries a warning for conservation. Soluble salts, including halite, syngenite and secondary gypsum, pervade the mortars. The gypsum largely results from the dissolution and recrystallization of older gypsum binders migrating through the masonry; the halite, at an inland site, most likely rose capillary-ward from saline groundwater or came from unwashed, evaporite-rich local sands; and syngenite forms when gypsum reacts with potassium from atmospheric deposition or fertilizers. These salts crystallize and swell with humidity, and their behavior directly explains the poor state of preservation visible across the complex today. Any modern repair mortar that ignored this chemistry would accelerate rather than arrest the decay.

In the end, the Winter Baths emerge as a case study in how Roman builders married imperial engineering knowledge to local geology, balancing logistics against performance with a precision that thermal modeling has only now made legible. By treating mortars not merely as glue but as chronological and technological documents, the research provides both a corrected timeline for the monument’s evolution and a rigorous template for designing compatible restoration materials, one based on the very recipes Roman craftsmen mixed nearly eighteen centuries ago. For a site whose excavated area still covers only a fraction of its ancient footprint, the stones and mortars of Thuburbo Majus are proving to be its most eloquent historians.

Subject of Research: Archaeometric characterization of Roman building stones and mortars from the Winter Baths of Thuburbo Majus, Tunisia

Article Title: Multi-analytical characterization of construction materials from the ‘Winter Baths’ of Thuburbo Majus (Tunisia): Implications for architectural chronology and conservation

Article References: Zaddem, A., Zoghlami, K., Fort, R., Romdhane, H. B., Coralini, A., & Gasmi, M. (2026). Multi-analytical characterization of construction materials from the ‘Winter Baths’ of Thuburbo Majus (Tunisia): Implications for architectural chronology and conservation. Archaeological and Anthropological Sciences, 18(10), Article 205. https://doi.org/10.1007/s12520-026-02563-w

Image Credits: AI Generated

DOI: 10.1007/s12520-026-02563-w

Keywords: Thuburbo Majus, Roman baths, archaeometry, mortars, hydraulic lime, cocciopesto, petrography, X-ray diffraction, Tunisia, conservation, building chronology, gypsum mortar