Sample Abstracts

The Yamuna River provides drinking water for over 57 million people across northern India, and microplastic contamination of freshwater has been linked to endocrine disruption in mammalian models.¹ While national-level surveys of the Ganga exist, no published study has mapped microplastic distribution along the Yamuna's 22-km stretch through Gurgaon, or tested whether concentrations track known industrial discharge points.² This project asked whether microplastic particle counts in the Yamuna correlate with proximity to identified industrial discharge points across that stretch.³

Water was collected in triplicate at five sites, spaced from 80 m to 21 km downstream of two mapped industrial outfalls, on three separate dates between June and August 2026. Samples were filtered through 20-µm nylon mesh and residue was examined at 40× under a stereo microscope by a blinded counter, with five non-overlapping fields per filter.⁴

Mean particle counts ranged from 63 fragments per 5-field sweep at 80 m to 8 fragments at 21 km, with a Spearman correlation of −0.91 (p = 0.004, n = 15) between distance and count. Site 3 (closest to Outfall B) showed a 6.1-fold higher load than Site 5.⁵ The strong distance-dependent gradient supports targeted enforcement of industrial discharge norms as a lever for reducing downstream microplastic load, and offers a low-cost school-lab protocol adaptable to other Indian urban rivers.⁶

Dengue cases in the National Capital Region rose 34% between 2019 and 2024, straining municipal health infrastructure during post-monsoon peaks.¹ Existing early-warning models rely on national temperature and humidity averages and do not account for the delayed vector-population response to short-duration heavy rainfall events, which are increasing under climate change.² This project tested whether a two-week lagged model using station-level daily rainfall could predict weekly dengue case counts in Gurgaon better than the standard temperature-humidity baseline.³

Publicly available weekly dengue case data from the MCG health portal (2019 to 2024, n = 260 weeks) was aligned with daily rainfall from IMD station GRG-01. Two Poisson regression models were fitted: a baseline using weekly mean temperature and humidity, and an extended model adding cumulative rainfall from the 8- to 22-day preceding window. Model comparison used AIC and 5-fold cross-validated RMSE.⁴

The rainfall-windowed model reduced cross-validated RMSE by 28% over the baseline (12.4 vs 17.2 cases per week), and correctly anticipated four of the five case peaks ≥ 50 cases at a two-week lead. False positives fell from 11 to 3 across the study period.⁵ Adding a rainfall-window term to existing dengue models is computationally trivial and could give municipal health teams a fortnight of additional preparation time before major outbreaks.⁶

Roughly 44% of rural Indian households consume unpasteurised milk, contributing to an estimated 500,000 cases per year of foodborne enteric illness in children under five.¹ Commercial batch pasteurisers cost ₹15,000 or more and require reliable grid power, both non-starters for a household budget of ₹80 per day and a village with four-hour daily outages.² The goal was to design and build a solar concentrator capable of pasteurising one litre of buffalo milk at 63 °C for 30 minutes, using materials totalling under ₹400 and no grid electricity.³

A compound parabolic concentrator was constructed from a discarded aluminium foil laminate and pinewood frame with an acceptance half-angle of 25° and a concentration ratio of 2.4. A 1-litre steel tiffin coated with matte black paint served as the receiver. Temperature was logged every 30 s with a DS18B20 sensor and Arduino Nano over 24 test days across three weather conditions in Meerut (Nov 2025 to Feb 2026).⁴

The receiver held ≥ 63 °C for 30 minutes on 18 of 24 test days (75%), including all 12 clear-sky days. Bill of materials totalled ₹376. Total-viable-count colony assays on treated versus untreated milk (n = 6) showed a mean reduction of 3.4 log CFU/mL.⁵ The design meets the WHO pasteurisation standard on typical winter days at less than 3% of commercial cost, offering a viable pathway for a village-level intervention pending independent field trials.⁶

Rural Indian primary classrooms average 42 students per teacher, well above the RTE-mandated 30, forcing single-file bench rows that limit peer interaction.¹ While collaborative learning is well studied in Western contexts, published evidence from the Indian primary sector on whether simple rearrangements alone (without additional teacher training) improve outcomes is limited.² This project tested whether rotating from front-facing rows to a facing-pair layout, without any other change to curriculum or teacher instruction, improves Class 5 mathematics scores over a 10-week term in a Government Primary School in Karnal district.³

With prior written parental consent and IRB approval from the Society for Science, 62 Class 5 students in two classrooms were randomly assigned by classroom to control (front-facing rows) or intervention (facing-pair) arrangements. Weekly 20-item mathematics quizzes were administered by an independent volunteer blinded to condition. Baseline was measured over the first two weeks. Analysis used a mixed-effects model with student as random effect.⁴

The intervention group improved by a mean of 3.8 marks (SD 4.1) over 10 weeks versus 1.6 (SD 3.9) in the control, a difference of 2.2 marks (95% CI 0.5 to 3.9; p = 0.011). Reported peer-explanation events per period rose from a mean of 0.7 to 4.2.⁵ A cost-free classroom rearrangement produced a small but statistically significant improvement in mathematics scores; the intervention's low cost and immediate feasibility make it a strong candidate for wider trial across the state's rural primary system.⁶