This page shows everything in the Lab Valley that a student can investigate: every live lab on the lab bench and the Mina Padi Rescue mission. It changes nothing in Sparkuna and reads no student data.
Nothing on this page is typed by hand. Every line a student reads, every right answer and every model comes from generated copies of the real files (Students Hub/partials/lab.js, docs/students-hub/missions/mina-padi.json, Central Hub/functions/paddy-model.js, the model card, the zone rules), and the screenshots are taken from the real student page. See Where this page comes from.
Start here: how to follow along
Sparkuna is Eduversal’s practice website for students. Across a bridge from each student’s island is the Lab Valley. On its lab bench are four short experiments (we call them labs). Each lab has the same four steps: predict, test, conclude, finish. Every lab is written five times, once for each grade band, in easier or harder words. A mission is a longer investigation with eight steps; the first one, Mina Padi Rescue, is open to students in Grades 7-12. Please try it and send us your notes.
On a big screen we open each lab with this page: the screenshots show every step, and the band cards show what Grade 1 and Grade 12 students read. You watch and ask.
About 30 minutesEveryone signs in at www.sparkuna.com (you are all testers, so every zone is open), goes to My Island → Lab Valley → lab bench, and runs the Leaf lab. Compare what you see with the cards on this page. Then use the small simulators here to test the models.
About 45 minutesEach person takes one lab and one band and checks the science and the English. The Biology member checks the mission’s model with the science check box. Notes go to the team’s WhatsApp group, and we fix them as they come.
One weekWords used on this page
- Lab (investigation)
- A short experiment on the lab bench: predict, test, conclude, finish. About 5 minutes.
- Grade band
- Students read text for their age: Grades 1-3, 4-6, 7-8, 9-10 and 11-12. Each lab has five versions.
- Predict
- The student chooses what they think will happen, before testing. There is no wrong prediction.
- Fair test
- Change one thing, keep everything else the same. The labs check this before the student can conclude.
- Conclusion question
- One question after the test, with three answers. One is right; the student then sees why.
- Model
- A small set of rules (a formula) that works out the result, like a simulation. It is simple and says it is a model.
- Mission
- A longer investigation for a person in the story (Pak Darto, a farmer). Eight steps, from brief to poster; Grades 9-12 add a ninth, a dry-season plan with no single right answer.
- Seed (two meanings)
- A garden seed is a reward the student plants. In the mission model, a weather seed is a number that fixes the weather of a trial, so a trial can be repeated exactly.
- mg/L
- Milligrams of oxygen dissolved in one litre of water. Fish struggle below about 3 mg/L.
- Zone
- One part of the Lab Valley. A zone opens at a Level, or after a number of careful practice runs.
Related pages: Island Garden + Lab Valley (the garden, paddy, kitchen and visitors), My Island, Run Engine (how practice runs work). For teachers and parents: the STEAM Integration guide (PDF) and the Lab Valley video.
Where the labs are
The Lab Valley has six zones. The first zone (the garden terraces) is open for everyone. The others open when a student reaches a Level or finishes enough careful practice runs, whichever comes first. Each lab lives in a zone, but since 18 Oct a lab is opened by the student’s lab chain, not by its zone: finish one lab and the next opens.
| Zone | Opens at | Labs |
|---|
The lab bench: the same four steps in every lab
The student taps the small table on the terraces. A list of labs opens. In each lab the student predicts, tests fairly, answers one conclusion question and finishes. The first finish of each lab gives a few garden seeds. Labs never give or take points and are not part of school grades.
- Open the lab bench
On the Lab Valley, tap the lab table on the garden terraces. Each lab shows its first sentence and a Start button (or ✓ Again when already finished). Since 18 Oct the labs come one after another: the next lab in the student’s chain shows its name with a lock (“Finish … to open it”) and opens when the one before is finished. See Chains and coverage.
- Predict
One question with three answers. Any answer moves on: a prediction is not marked right or wrong. It is shown again beside the conclusion (“You predicted … Did your results agree?”).
- Test
The student changes the settings and reads a chart. The “I have my results” button appears only after a fair test (each lab’s rule is below).
- Conclude
One question with three answers, in a new order every time. After answering, the right answer is marked and a short “why” is shown, right or wrong.
- Finish
“Finish the investigation” saves it in the student’s lab notebook. The first finish of each lab gives 2 garden seeds (never more than 15 in hand). After that it is just for learning.
What the student sees
What is recorded
When a student finishes a lab, the page sends three things: which lab, whether the conclusion answer was right, and whether a prediction was made. The server keeps, for each lab, how many times it was finished, how many of those were reported, how many had the right conclusion and how many had a prediction (garden.js labLog). These numbers are for HQ only (Usage Insights). They are never a grade and never change the seeds.
The rules this page explains, checked against the student page
| Rule | Still in the student page? |
|---|
Chains and curriculum coverage
Schools teach at different speeds, so the labs do not follow one school’s pacing. Instead each grade band has its own chain: the first lab is open, and finishing a lab opens the next one. Labs further on show their name with a lock. Testers and HQ observers see every lab open. A lab finished before 18 Oct stays open and opens the one after it.
The coverage map shows which parts of each band’s syllabus the labs touch, so we can see what is still missing. The aim is that, over time, the labs touch almost every topic.
The chain
Coverage of the syllabus
Units: Primary and Lower Secondary = stage × sub-strand (Central Hub resources/Primary-Science.json, Checkpoint-Science.json; Lower Secondary Stages 7-9 are taught in Grades 7-8); IGCSE and AS & A Level = syllabus sections (resources/IGCSE-*.json, AS-A-Level-*.json). A unit counts as touched when a lab lists one of its objective codes for that band (partials/lab.js curriculum). Skills strands (Thinking and Working Scientifically, Science in Context) are not counted. Chains: Central Hub/functions/garden.js LAB_CHAINS.
A water plant sits in a beaker under a lamp. The plant makes its food by photosynthesis and gives out oxygen as small bubbles. The student moves the lamp nearer or further, adds baking soda (it gives carbon dioxide) and warms the water, then counts the bubbles for one minute. More bubbles = faster photosynthesis.
Grades 1-3 change only the lamp; the soda and the water stay at the best setting.
The model behind it
- Light falls with the square of the distance from the lamp (inverse square law): at 20 cm the light is ¼ of the light at 10 cm. The model takes 10 cm as full light.
- Carbon dioxide comes from the pinches of sodium hydrogen carbonate (none, 1, 2 or 3).
- Temperature: enzymes work best near 30 °C and are damaged above about 40 °C.
- Limiting factor: the rate is set by whichever of light or carbon dioxide is in shorter supply, times the temperature effect. Up to about 60 bubbles a minute.
- Real counts are never exact, so each count wobbles by up to 2 bubbles (the same setting counted again can give a slightly different number).
The fair test
At least 3 counts, and two of them at different lamp distances with the same soda and the same water temperature. Only then does “I have my results” appear. Grades 9-12 also get a tip: try the lamp close with no soda, and very warm water.
Step by step
Try the model
The chart shows a count at every lamp distance with these settings. It runs bubbles() from lab.js.
Bubbles per minute at each lamp distance
Every band’s text
Exactly as students read it. The right conclusion answer is ticked (students see the three answers in a shuffled order).
The model code (exact copy of lab.js)
A picture of the nitrogen cycle with seven steps. The student taps the steps one by one and reads what happens at each (in their band’s words). Then they look at the soil nitrogen number of their own garden beds: harvests use nitrogen, peanuts and compost put it back.
The model behind it
The cycle itself is a diagram (seven steps, eight arrows), not a calculation. The numbers come from the student’s own beds, with the garden rules below (garden.js SOIL). Soil changes only at a harvest or with compost, and nothing a student has is ever lost.
The fair test
This lab has no settings to change. The student must tap at least 5 of the 7 steps; then “I followed the cycle” appears.
Step by step
The seven steps of the cycle, in every band
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
The student’s own rice-fish paddy, over one day. In daylight, algae and water plants under the water make oxygen. Fish, algae and tiny living things use oxygen all the time, day and night. The line shows how much oxygen is in the water each hour. It is lowest just before sunrise.
The model behind it
- Gain: photosynthesis under the water while the sun is up (06:00-18:00). Less on a cloudy day. Rice leaves stand above the water, so their oxygen goes to the air, and they shade the algae a little.
- Loss: respiration of the fish, the algae and the microbes that break down waste, day and night. More fish use more. Rice roots take up much of the fish waste, so less waste is left to break down.
- Air: oxygen moves between the water and the air, towards 7.8 mg/L (full, at about 28 °C). Fresh stream water does this faster.
- Same model as the mission (since 17 Oct 2026): the lab runs paddy model v2, the one Mina Padi Rescue uses, so the two never disagree. It shows day 7 of a batch (the fish half grown) on one fixed weather pattern: the same paddy history, then “sunny since yesterday” or, with Cloudy day, “cloudy since yesterday”, so the chart shows both the lower midday peak and the lower dawn after cloudy weather. On that day 3 fish stay above 3 mg/L (lowest 4.6), 5 fish touch it (3.1), 8 fish fall below (2.0; 0.9 when cloudy). The lowest point always falls around dawn.
The fair test
The grey dashed line is the student’s own paddy today (its fish and rice, sunny, still water). The student must try 3 set-ups that each change exactly ONE thing from it. If a set-up changes two things, the page says “you cannot tell which one made the difference”.
Step by step
Try the model
Runs oxygenDay() and draws with oxygenChart() from lab.js, exactly as students see it.
Oxygen in the paddy over one day
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Bees carry pollen from flower to flower; a flower with pollen can become a fruit. The student changes the orchard (more kinds of flowers, a bee hotel, insecticide spray, a rainy morning) and compares the bee visits and how many of 20 flowers become fruits.
The model behind it
- Bee visits rise with the kinds of flowering plants nearby (food through more of the year), but level off.
- Insecticide on open flowers kills or drives away the bees (visits × ¼). A bee hotel (nest sites) adds bees (× 1.3). Bees hardly fly in rain (× 0.35).
- Fruits from 20 flowers rise with visits but saturate: once most flowers have pollen, more visits add little (diminishing returns).
- The student’s own orchard starts from the kinds of plants growing in their beds now (1 to 5). Model units, not field data.
The fair test
The same rule as the oxygen lab: 3 set-ups that each change exactly ONE thing from the student’s own orchard.
Step by step
Try the model
Runs pollination() and pollinatorChart() from lab.js.
Bee visits and fruits
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
A solar panel in the Energy Yard runs fans. The student tilts it, makes the day cloudy, lets dust settle on it or adds an air gap below it, and compares the power at midday and the fans it can run. Grades 7-12 also see the light reaching the panel and its efficiency.
The model behind it
- Light reaching the panel = intensity × area × cos θ, with the midday sun 10° from overhead (Indonesia is near the equator). Sunny 1000 W/m², cloudy 300 W/m². The panel is 1.6 m².
- Efficiency is 20% at 25 °C and falls by 0.4% for every °C above it (silicon panels: about 0.3-0.5%). The panel runs at 65 °C flat on a roof, 55 °C with an air gap (about 10 °C cooler), 35 °C on a cloudy day.
- Dust after weeks without rain blocks a quarter of the light. Fans use 40 W each.
- Starting set-up: tilted 15°, sunny, clean, no air gap (268 W, 6 fans). Model, not a datasheet.
Curriculum links
The fair test
The pollinator lab’s rule: 3 set-ups that each change exactly ONE thing from the starting set-up.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
A belt of mangrove forest stands between the sea and a village. The student changes its width, makes a storm or a high tide, or swaps the grown forest for a new planting, and compares the waves that reach the village with the young fish sheltering in the roots.
The model behind it
- Wave height falls by the same fraction for every metre of forest: H = H₀·e−kx, k = 0.005 per m (about 40% less per 100 m). Field studies report 13-66% per 100 m (McIvor et al. 2012, Reduction of wind and swell waves by mangroves, The Nature Conservancy and Wetlands International).
- Waves are 50 cm on a normal day and 200 cm in a storm. A high tide covers more of the roots (k × 0.6). A new planting has thin trunks and few prop roots (k × 0.4).
- Young fish rise with the width and level off (model units, not field counts); a new planting shelters far fewer (× 0.3).
- Starting set-up: 100 m of grown forest, normal waves, low tide (30 cm waves, 34 young fish).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the starting coast.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
The student carries a weather station up the volcano (0 to 3000 m), moves it to the rain-shadow side or switches to the dry season, and compares the midday temperature with the rain in a typical month.
The model behind it
- Temperature falls about 6.5 °C for every 1000 m (the average lapse rate), from 31 °C at the coast at midday. The rain-shadow side is about 2 °C warmer (air coming down warms: the foehn effect); the dry season 1 °C warmer (fewer clouds).
- Rain on the side facing the wind rises up the slope to a peak near 1500 m, then falls (less water is left in the air higher up): 300 mm a month at the coast, about 900 mm at 1500 m in the wet season. The rain-shadow side gets 45% of it, a dry-season month 15%.
- Typical months, not one year’s data. Starting set-up: 1000 m, wind side, wet season (25 °C, 860 mm).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the lookout station. Every other height counts as one change.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1 (19 October 2026). An upland rice terrace is too acid. The student spreads lime (calcium carbonate) on it, changes the amount, grinds it to a powder, tries a hot week or acid rain, and compares the soil pH after a week with the rice from the bed. Grades 7-12 also see how much of the lime reacted.
The model behind it
- Acid: the soil holds 6 units of acid (about pH 4.8); acid rain adds 4. Neutralisation: CaCO₃ + 2H⁺ → Ca²⁺ + H₂O + CO₂, 1 kg of lime ≈ 2 units.
- Rate: in one week only part of the lime reacts, f = 1 − e−k; k is 4 × bigger for powder (surface area) and 2 × for a hot week (35 °C instead of 25 °C: the rule of thumb of twice the rate for 10 °C). Chips: 50% reacts in a week, powder 94%.
- pH = 7 − 2.6·log₁₀(1 + acid left). Extra carbonate is a weak, nearly insoluble base, so it cannot push the soil much past pH 8: 7 + 1.2(1 − e−x/4).
- Rice does best near pH 5.5-6.5: kg from the bed = 6·e−((pH − 6.3)/1.5)², so too much lime is also bad. Starting set-up: 2 kg of chips, a normal week (pH 5.2, 3.4 kg of rice). Model, not soil-lab data.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the sour terrace.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1. An iron gate in the Energy Yard, painted, with a scratch. The student swaps the paint for bare iron or a zinc coat, adds salty sea spray or a roof, or mends the scratch, and compares the rust and the iron eaten away after two years. It shows why paint (a barrier) fails at a scratch and zinc (a barrier that is also sacrificial) does not.
The model behind it
- Bare steel loses 40 µm a year in the humid tropics inland, 3 × that in salty sea spray and a quarter under a roof (ISO 9223 corrosivity categories C3, C5 and C2 as the guide; model units), a little less each year as rust builds up: loss = r·t0.8.
- Paint is a barrier only where it is whole: at the scratch the iron rusts as deep as bare iron and the rust creeps under the paint. Whole paint lasts about 4 years (1.6 by the sea).
- Zinc (85 µm of galvanising) is oxidised instead of the iron, even at the scratch, at about 1/25 of the iron's rate, so it lasts for decades here.
- Starting set-up: painted gate with a scratch, inland, in the open (25% of the gate rusty, 70 µm deep at the scratch).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the painted, scratched gate.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1. A bucket of water from the paddy channel under a turbidity meter. The student adds mud, filters it, boils it or adds chlorine, and compares the light getting through the sample with the germs left in it. Its one message: clear is not the same as safe. Every screen says never to drink paddy or river water.
The model behind it
- Cloudiness: each spoon of mud adds 40 NTU. Light through falls by the same fraction for each extra particle: I = I₀e−0.012·NTU (Beer-Lambert; for Grades 11-12 the same law as X-ray attenuation). 1 spoon lets through 62%, 2 spoons 38%.
- Germs (model units per drop) come from waste upstream and do not show on the meter. A sand-and-cloth filter takes out 90% of the mud but only 70% of the germs; boiling kills them all.
- Chlorine kills 99.9% in clear water (under 10 NTU) but only 95% in muddy water, where the particles shelter germs: the reason water works filter before they add chlorine.
- Starting set-up: 2 spoons of mud, untreated (80 NTU, 38% of the light, 400 germs a drop).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the paddy water.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1. Amylase, the enzyme in saliva, with cooked rice in the island kitchen. The student changes the temperature, adds lime juice, leaves it 20 minutes instead of 5 or boils the enzyme first, and compares the starch turned to sugar with the share of the enzyme still working. Grades 4-12 see the iodine colour. The plan called it "an enzyme in rice"; the name now says what happens, "starch to sugar".
The model behind it
- Rate doubles for every 10 °C while the enzyme works (2(T − 37)/10); lime juice (pH 3) cuts it to 5%.
- Denaturation: the share still working after t minutes is e−kd·t, kd = 0.02·e(T − 37)/3.5 a minute: almost none lost at 30 °C, a fifth in 5 minutes at 40 °C, nearly all at 50 °C, all at 60 °C. A boiled enzyme does nothing.
- Starch turned to sugar = 1 − e−0.195 × rate × minutes of working enzyme. So the best temperature depends on how long the test runs: 40 °C for 5 minutes; at 20 minutes 30 °C almost catches up and 50 °C falls far behind (the Grades 11-12 question).
- Starting set-up: 30 °C, 5 minutes (45% of the starch turned to sugar, 99% of the enzyme still working).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the kitchen set-up.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1. A leafy shoot in a potometer. The student changes the air temperature, puts it in the dark, turns on a fan or makes the air damp after rain, and compares how far the bubble moves in 5 minutes with how open the stomata are.
The model behind it
- Transpiration ∝ the water-vapour deficit of the air = saturated vapour pressure at T (Tetens' formula, 0.611·e17.27T/(T + 237.3) kPa) × (1 − relative humidity); 60% normally, 90% after rain.
- Wind × 1.8 (it carries away the damp air around the leaf). Stomata are 90% open in light, 10% in the dark.
- The bubble moves 13.1 mm per kPa in 5 minutes with open stomata. From 30 °C to 35 °C at the same humidity the rate rises by about a third.
- Starting set-up: 30 °C, sunny, still, normal air (22 mm in 5 minutes).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the shoot on a sunny day.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1. A farmer keeps seed from the best rice plants for the next generation. The student changes what share is kept, tries a field with uneven soil or ten generations instead of five, and compares the rice per plant at the end with the gain each generation. The copy says "selection" and "next generation" (Sparkuna words rule).
The model behind it
- The field's plants give 20 g of rice each on average, standard deviation 4 g. Keeping the best p% for seed, the kept plants beat the average by S = i·σ (selection intensity i of a normal distribution: 2.06 / 1.75 / 1.27 / 0.80 / 0 for 5 / 10 / 25 / 50 / 100%).
- Breeder's equation: each generation gains R = h²S, heritability h² = 0.4 on even soil and 0.15 on uneven soil (most differences then come from the soil, not the genes).
- Variance is kept the same each generation (a simple model: real selection also uses up variety).
- Starting set-up: keep the best 25%, even soil, five generations (S = 5.1 g, R = 2 g a generation, 30 g per plant at the end).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the farmer's plan.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1. A cable from the yard's 24 V supply to 12 classroom fans. The student changes the length, uses a thick cable or aluminium instead of copper, or adds transformers that step the voltage up for the cable, and compares the power reaching the fans with the power wasted in the cable as heat.
The model behind it
- Cable resistance R = ρ·2L/A (there and back): copper ρ = 1.7 × 10⁻⁸ Ω m, aluminium 2.8 × 10⁻⁸; 1.5 mm² or 4 mm².
- The fans: 12 × 40 W at 24 V, so 1.2 Ω together. Current I = 24 / (1.2 + R); the fans get I²·1.2 and the cable wastes I²R.
- Transformers (ideal, VpIp = VsIs) step 24 V up × 10 for the cable and down again at the fans: the cable carries a tenth of the current, so it acts like R/100 (120 W wasted → 4 W).
- Starting set-up: 50 m of 1.5 mm² copper (1.13 Ω; 127 W to the fans, 120 W wasted, 3 fans running).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the 50 m copper cable.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 1. A classroom under a dark metal roof at midday. The student adds insulation, paints the roof white, swaps it for shiny bare metal, plants shade trees or opens windows for a cross-breeze, and compares how much warmer the classroom is than the air outside with the roof temperature.
The model behind it
- The roof warms until it loses energy as fast as it absorbs it: absorptivity × sunlight = 12·(Troof − Tair) + emissivity·σ·(Troof⁴ − Tsky⁴), with air 32 °C, sky 10 °C colder, sunshine 1000 W/m² (400 under shade trees).
- Surfaces: dark paint absorbs 0.9 and emits 0.9 (roof 76 °C); white paint 0.25 / 0.9 (43 °C); shiny bare metal 0.35 / 0.1 (59 °C: it reflects well but emits badly).
- The room = the air + 2 °C (30 people) + 0.08 × (Troof − Tair) ÷ (1 + cm of insulation / 2.5); a cross-breeze halves what comes through the roof and takes off 1 °C more.
- Starting set-up: dark roof, no insulation, no trees, windows half open (classroom 37.5 °C, 5.5 °C above outside). Model, not a building simulation.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the classroom under a dark roof.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2 (22 October 2026). One steel string of a sasando, the palm-leaf harp from Rote Island. The student changes the length, tightens it, swaps it for a thicker string, plucks harder or listens from 4 m away, and compares the pitch (Hz) with the loudness (dB). Length, tension and thickness move the pitch; the pluck and the distance move only the loudness.
The model behind it
- The fundamental of a string fixed at both ends: f = (1/2L)·√(T/μ), with μ = steel density 7850 kg/m³ × the cross-section of a 0.3 mm (or 0.5 mm) wire; tension 60 N, or 120 N tightened.
- Loudness at 1 m: 60 dB for a soft pluck; a harder pluck (three times the amplitude) +9.5 dB; 4 m away −12 dB (6 dB less for each doubling of the distance).
- Starting set-up: 40 cm, 60 N, 0.3 mm, soft pluck, 1 m (411 Hz, 60 dB; wave speed on the string 329 m/s). Half the length doubles the frequency; twice the tension multiplies it by √2.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the 40 cm string.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2. The tide at the mangrove over the Moon's month. The student moves through the days from new moon to full moon, puts the Moon at its closest to Earth or adds a strong wind from the sea, and compares the high tide with the tide range. Spring tides come at new and full moon, neap tides at the half moons.
The model behind it
- Two tides added: the Moon's 50 cm and the Sun's 23 cm (0.46 of the Moon's, because the tide-raising effect goes as M/r³). Height = √(M² + S² + 2MS·cos(4π·day/29.5)): 73 cm in a line, 27 cm at right angles.
- Perigee (363 300 km, not 384 400): the Moon's tide × (384.4/363.3)³ = 1.18. Wind from the sea piles the water 30 cm higher without changing the daily range.
- Starting set-up: half moon, day 7 (high tide 28 cm above the average sea, range 55 cm). A simple twice-a-day tide: real Indonesian coasts also have one-a-day tides.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the half moon.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2. A fisher's outrigger canoe (jukung). The student loads cargo, makes the hull wider, takes it into fresh river water or builds the hull of heavy ironwood (ulin), and compares how deep it sits with how much side is left above the water. At 400 kg of cargo the water comes over the side.
The model behind it
- Archimedes: the canoe sinks until it pushes aside its own mass of water: depth = total mass ÷ (water density × waterline area). Waterline area = 5 m × width (0.6 m or 0.8 m) × 0.55.
- Masses: hull 80 kg (light wood) or 160 kg (ironwood), fisher 60 kg, cargo 0-400 kg. Sea water 1025 kg/m³, river water 1000. The side is 30 cm high.
- Starting set-up: 100 kg of cargo, 0.6 m hull, light wood, sea water (240 kg, 14.2 cm deep, 15.8 cm of side left; water pressure under the hull 1.4 kPa). 100 kg more ≈ 6 cm deeper.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the jukung with 100 kg of fish.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2. A sack of potatoes sliding down a harvest slide on the volcano slopes. The student changes the slope, swaps rough wood for smooth bamboo, uses a sack twice as heavy or a slide twice as long, and compares the speed at the bottom with the time it takes. The heavy sack arrives at exactly the same time.
The model behind it
- Along the slope: a = g(sin θ − μ cos θ), sacking on rough wood μ = 0.45, on polished bamboo 0.2; no air resistance (small at these speeds). From rest v = √(2aL), t = v/a.
- The mass cancels: twice the mass, twice the pull and twice the friction. Energy: the share of the potential energy turned to heat by friction = μ cos θ / sin θ.
- Starting set-up: 30°, rough wood, 25 kg, 20 m (a = 1.08 m/s², 6.6 m/s at the bottom after 6.1 s; 78% of ΔEp becomes heat). At 25° friction almost balances gravity (a = 0.14 m/s²).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the 20 m wooden slide at 30°.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2. 100 g of fuel burned under a pot of 2 kg of water. The student switches between wood, charcoal, biogas and LPG, puts a lid on the pot, builds a wind shield or half closes the air vents, and compares the heat reaching the water with the carbon monoxide made. Charcoal looks clean but makes the most carbon monoxide.
The model behind it
- Energy stored (typical, MJ/kg): air-dry wood 15, charcoal 29, biogas 19, LPG 46. Share reaching the water: open wood fire 15%, charcoal stove 25%, biogas and LPG burners 50%; a lid × 1.3, a wind shield × 1.25.
- Air vents half closed: incomplete combustion releases 80% of the energy and 4 × the carbon monoxide. CO per 100 g (typical cookstove tests): wood 6 g, charcoal 15 g, biogas 0.3 g, LPG 0.2 g.
- Starting set-up: wood, no lid, no shield, vents open (225 kJ into the water, 25 → 52 °C by q = mcΔT; 6 g of CO). Every screen says never to burn charcoal indoors.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the open wood fire.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2. Urea spread on a flooded rice field, one week. The student changes the pH of the paddy water, makes the water hot or the week windy, or presses the urea into the mud, and compares the nitrogen lost to the air as ammonia with the nitrogen kept for the rice.
The model behind it
- Equilibrium: NH₄⁺ ⇌ NH₃ + H⁺, pKa = 0.09018 + 2729.92/T (Emerson et al. 1975: 9.16 at 28 °C, 8.95 at 35 °C); share of NH₃ = 1/(1 + 10^(pKa − pH)).
- A week of losses: ammonia leaves at 0.8 × share a day (× 1.5 windy, × 0.1 pressed into the mud), the rice and soil take up 0.15 a day; lost = kv/(kv + kt)·(1 − e^(−7(kv + kt))). Losses run from 2% (pH 7) to about 80% (pH 9.5), the range measured in flooded rice.
- Starting set-up: pH 8, 28 °C, a still week, on the water (6.5% of the ammonium is NH₃; 20% of the nitrogen lost, 80% kept). Pressed into the mud: about 2% lost.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from urea thrown on the paddy water.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2. Splitting water with electricity from the yard's solar panel, to store the energy as hydrogen. The student changes the current, runs it for 2 hours, fills the cell with salt water or with pure water, and compares the gas collected at each electrode. Brine gives chlorine at the + electrode; pure water hardly conducts.
The model behind it
- Faraday: Q = It, n(e⁻) = Q/F (F = 96 500 C/mol); 2 e⁻ per H₂ and 4 e⁻ per O₂, so twice the volume of hydrogen; 24 litres a mole. An ideal cell: every electron makes gas.
- The liquid: water with a little sodium sulfate (conducts); pure water lets only 2% of the current through; concentrated brine gives chlorine at the anode (2 e⁻ per Cl₂, none dissolving).
- Starting set-up: 2 A for 1 hour, sodium sulfate solution (7200 C, 0.075 mol of electrons: 0.9 litres of hydrogen and 0.45 litres of oxygen).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the cell at 2 A for 1 hour.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 2. A paper chromatogram of a green batik dye made of three model dyes. The student changes the solvent, puts on a big blob instead of a dot, draws the start line in ink, lets the solvent come over the start line or stops the run early, and compares how many colours come apart. Only the water + alcohol mixture separates all three.
The model behind it
- Rf values (made up, per solvent: blue / brown / yellow): water 0 / 0.45 / 0.55; water + alcohol 0.35 / 0.55 / 0.80; alcohol 0.70 / 0.15 / 0.78. The solvent runs 80 mm (40 if stopped early).
- Spots are as wide as the dot put on (3 mm, or 11 mm for a blob) + 15% of the distance moved. Neighbours are seen apart when their centres are further apart than their average width. An ink line adds its own purple spot (Rf 0.2); solvent over the line washes the dye into the jar.
- Starting set-up: water + alcohol, a 3 mm dot, pencil line, 80 mm run (3 colours apart, smallest gap 8 mm).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the green batik dye in water + alcohol.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. A 1 m pole on flat ground in Pontianak, a city on the Equator. The student moves the time of day, changes the date to 21 June, moves the pole to Jakarta or makes it 2 m tall, and compares the shadow’s length and direction and the Sun’s height. At noon on 21 March the Sun is straight overhead: no shadow at all.
The model behind it
- Sun’s height h: sin h = sin φ sin δ + cos φ cos δ cos H, with the hour angle H = 15° for each hour from noon (sun time). Latitude φ = 0° (Pontianak) or 6.2° S (Jakarta); the Sun’s declination δ = 0° on 21 March, +23.44° on 21 June.
- Shadow = pole height ÷ tan h, pointing away from the Sun (west in the morning, east in the afternoon, south at noon in June and in Jakarta).
- Starting set-up: the 1 m pole in Pontianak at noon on 21 March (shadow 0 cm, Sun 90° up). Grades 7-8 carry skills codes only: Checkpoint Stages 7-9 have no outcome on the Sun’s daily path.
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the 1 m pole at noon on 21 March in Pontianak.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. The lamps of a night-market stall on a battery. The student changes the number of lamps, joins them in parallel instead of in series, uses four cells instead of two or adds a long thin cable, and compares each lamp’s brightness and the battery life. Six lamps in parallel blow the 1 A fuse.
The model behind it
- Cells 1.5 V, internal resistance 0.4 Ω each; lamps 12 Ω (made for 3 V, 0.75 W = 100%), treated as fixed resistors. Series R = nR, parallel R = R/n; a long thin cable adds 3 Ω.
- Current I = E ÷ (R + r); brightness = each lamp’s I²R ÷ 0.75 W; battery life = 2 Ah ÷ I. The fuse opens above 1 A.
- Starting set-up: three lamps in series on two cells (11% brightness, 24.5 hours).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from three lamps in series on two cells.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. A small water-wheel generator for a village. The student changes the number of turns on the coil, uses a stronger magnet, opens the sluice so it spins twice as fast or stops the water, and compares the peak voltage and the frequency. Only the speed changes the frequency; a still wheel gives nothing.
The model behind it
- Faraday’s law: peak e.m.f. E₀ = NBAω for a 10 cm × 10 cm coil (A = 0.01 m²) in 0.2 T (or 0.5 T), turning 10 times a second (20 with the sluice open), ω = 2π × turns a second; frequency = turns a second.
- A 12 V lamp lights fully at 12 V r.m.s. (E₀ ÷ √2). No losses inside the coil.
- Starting set-up: 100 turns, 0.2 T, 10 turns a second (peak 12.6 V, 10 Hz).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the 100-turn generator in a gentle stream.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. A fish 50 cm under calm water. The student changes the angle they look from, puts the fish 1 m down, puts on polarised sunglasses or comes back at night with a lamp, and compares where the fish seems to be and the glare off the water. The glasses change the glare, never where the fish seems.
The model behind it
- Refraction: sin i = 1.33 sin r. Along the line of sight the fish seems at depth d·tan r / tan i (d/1.33 straight down); a spear aimed at what you see passes d(tan i − tan r) beyond it.
- Glare = sky light reflected by the water (Fresnel equations, unpolarised): 2% straight down, 13% at 70°. Polarised sunglasses pass only the p-part, which is zero at Brewster’s angle (53°), and halve the fish’s light. At night there is no sky glare.
- Starting set-up: the fish 50 cm down, seen at 40° (seems 33 cm deep, glare 2.4%).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the fish 50 cm down, seen at 40°.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. A 14-year-old walking up the volcano path. The student changes the pace from sitting still to running up, gives them 8 weeks of training, takes them up to 2 000 m or adds a 10 kg pack, and compares the heart rate and the breaths a minute. A model of a typical teenager made from average values, not a medical tool.
The model behind it
- Oxygen used (mL/kg/min, 50 kg): sitting 3.5, slow walk 10, brisk walk 15, steep path 25, running up 40; a 10 kg pack × 1.2. Cardiac output Q = 4 + 6 × V̇O₂ (L/min), × 1.1 at 2 000 m.
- Heart rate = Q ÷ stroke volume (70 mL rising to 95 mL in hard work; × 1.25 after training). Above 70% of V̇O₂max (45, trained 55 mL/kg/min) lactic acid builds up and breathing rises × 1.35.
- Starting set-up: a brisk walk (103 beats and 30 breaths a minute).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from a 14-year-old on a brisk walk.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. Salting fish to make ikan asin. The student soaks a slice of fish for a day in brines from 0% to 26% salt, uses the same grams of sugar instead, cuts it 3 cm thick or keeps it at 37 °C, and compares the water drawn out and the germs after a day. The germs stop at about 15% salt; a thick slice spoils in the middle.
The model behind it
- Water potential of the brine ψ = −CRT × particles (salt gives two ions, osmotic coefficient 0.93; sugar about 1/11 of the particles for the same grams); fish tissue ≈ −0.75 MPa. Water lost after a day = 40% × (1 − e^(−Δψ/8 MPa)), reached with a time constant of 6 h (54 h for 3 cm).
- Germs (thousand per gram, from 10): doubling every 2 h at 30 °C (1 h at 37 °C), slowed by the water activity a_w ≈ 1 − 0.018 × osmolality and stopped below 0.91.
- Starting set-up: a 1 cm slice in 10% brine at 30 °C (25% of its water out, 127 thousand germs per gram).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from a 1 cm slice in 10% brine.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. Radon gas from the volcano’s soil, sealed in a jar next to a counter. The student waits some days, puts paper or 5 mm of aluminium in front of the counter or moves it 20 cm away, and compares the count rate and the radon atoms left. Shielding lowers the count but never changes the decay.
The model behind it
- Decay: radon-222, half-life 3.8 days, N = N₀e^(−λt). At the start 800 counts a minute from the jar (model shares: 60% α, 30% β, 10% γ) + 20 a minute background.
- Shielding: paper stops α; 5 mm of aluminium stops α and β; 20 cm away α cannot reach the tube and β and γ fall with the square of the distance (÷ 100).
- Starting set-up: the fresh jar, the counter 2 cm away (820 counts a minute).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from the fresh jar with the counter 2 cm away.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Batch 3. Turning palm oil into biodiesel. The student changes the methanol, warms the mixture, gives it 2 hours, uses oil with a little water in it or leaves out the catalyst, and compares the biodiesel and the soap made. Simulation only: the practical never uses methanol.
The model behind it
- Equilibrium per ester link: K = 3y²/((1 − y)(R − 3y)) = 8, R = moles of methanol per mole of oil (74% at 3 to 1, 91% at 6 to 1, 96% at 12 to 1).
- Rate: y = y_eq(1 − e^(−kt)), k = 0.06 per minute at 40 °C, × 2.7 at 60 °C (Eₐ ≈ 43 kJ/mol), × 1/1000 without the catalyst; water in the oil turns 70% of the catalyst into soap.
- Starting set-up: 6 to 1 methanol at 40 °C for 30 minutes (76% made, no soap).
Curriculum links
The fair test
The same rule: 3 set-ups that each change exactly ONE thing from 6 to 1 methanol at 40 °C for 30 minutes.
Step by step
Try the model
Every band’s text
Exactly as students read it. The right conclusion answer is ticked.
Mission
The student page (/lab?mission=mina-padi, entered from a gold row at the top of the lab bench) is live at every school, and the screenshots below are taken from it. It is for Grades 7-12 (and testers) and opens with the rice-fish paddy zone. The server runs every trial with the real model and checks the evidence the student gives; the first finish gives 4 garden seeds and a notebook stamp. Since 22 October 2026, Grades 9-12 add step 9, the dry-season plan (Grades 7-8 keep the eight steps). The subject team checks it while it is live: send your notes from the science check box and we fix them.
Pak Darto, a farmer, keeps fish in his rice field (mina padi). He wants more fish, but too many fish stay small. The student tests 3, 5 and 8 fish on a model of the paddy (10 days in a few seconds), finds the best number, tries ONE change that could help (rice, or fresh stream water), explains the advice with their own results and makes him a poster. Grades 9-12 then plan for the dry season: they make at least two plans (fish, rice, and how many days to take water from the stream that the fields below also need), compare them on five rows, and choose one, naming what matters most and what it costs. Different students can choose different plans and all be right, as long as their sentences fit their own numbers.
Curriculum map
The paddy model behind the mission
The mission runs on a small model of a rice-fish paddy: 1 to 8 fish, rice or no rice, still or fresh water (every day, or on some days in turns for the dry-season plan), sunny or cloudy days. Every quarter of an hour it works out the oxygen, the waste, the algae and how much each fish grows. The same plan with the same weather seed always gives the same result, so the server can check what a student reports. It is a teaching model, not field data.
Science check — Biology member of the subject team
The mission is already open to students, so your check makes it better while it is live. Please check the points below with the simulator further down, then fill in the note and send it in the team’s WhatsApp group (or to Alif). Any change you ask for is fixed, and the model card in the repository is updated with your name and result.
Nothing is saved on this page; the note is only copied for you to send.
Try the model (the same code as the mission)
Runs SH_PADDY_MODEL.run(), a generated copy of Central Hub/functions/paddy-model.js (version 2), for 10 days.
Oxygen every hour, all days
One day, hour by hour
Behaviour, worked out now on this page
Mean of 40 trials (weather seeds 1 to 40), 10 days. Compare it with the table the model was tuned to (in the model card below): the story should be the same.
Model parameters
All numbers in the model (SH_PADDY_MODEL.P)
Model card
Where this page comes from
A script copies the real files into this page. When a lab, the mission or the model changes, run the two commands below and the page shows the new version. Nobody types the student text here by hand.
- Text and models
node scripts/island/build-lab-valley-guide.js(add--checkto only check). It copieslab.jsandpaddy-model.jsas they are, reads the mission JSON, the model card and the zone rules, and stops if a lab-bench rule explained on this page is no longer in the student page. - Screenshots
node scripts/island/shoot-lab-valley-guide.jsopens the real student page on a test copy (no real data, no Firestore) as a Grade 8 and a Grade 2 student, walks every lab and saves the pictures. For the mission, the mission page’s own test copy saves full-page pictures and--mission-from=<folder>crops them to each step. Then run step 1 again. - Publish
Rebuild Central Hub (
node build.js) and ship it as usual.