Experimental Assumptions Catalogue

The dropper method is consistent enough to produce balls within ±15% diameter. Ten balls are used, so size variation averages out across the group.

Balls with larger surface area produce more CO₂ per unit time, inflating the signal. Some outlier balls (very small or non-spherical) will contribute proportionally less.

Adds scatter between groups but is unlikely to reverse condition rankings — size variation is random across conditions.

Large between-group differences for the same condition (e.g. one group's Purple ΔpH much higher) that can't be explained by other factors.

Organisms are prepared fresh for each session. Chlorella is cultured and checked for green colour; yeast is activated in warm water before use.

The culture's overall health is assumed but not measured. A compromised batch affects all groups equally — class results will all shift in the same direction.

If the batch is compromised, all signals will be smaller than expected. Rankings may still be preserved, but the magnitude of ΔpH will be lower. A dead batch gives zero signal in all conditions.

All ΔpH values substantially smaller than historical reference values across all groups and conditions simultaneously.

The alginate solution is continuously stirred during ball-making, keeping cells in suspension. Cell density is approximately uniform across balls made from the same batch.

Cells may settle slightly if stirring pauses. The effect is very small relative to the between-condition signal — it adds minor scatter but does not create systematic bias.

Minor variation in signal magnitude. With 10 balls per condition, statistical averaging makes this effectively negligible.

Cannot be directly detected from plate reader data alone. Would require counting cells per ball under microscopy.

Two rinses in fresh buffer are specified. Calcium alginate dissolves slowly, so surface residue is mostly removed with the first rinse; the second provides additional insurance.

Alginate releases CO₂ as it degrades in acidic conditions. Residual alginate could contribute a small baseline CO₂ signal on top of the biological signal.

All conditions receive a small positive CO₂ offset. This would slightly reduce ΔpH in photosynthesis and exaggerate ΔpH in respiration. Systematic across all conditions, so rankings are preserved.

Dark control ΔpH more negative than expected; T0 pH lower than the prepared buffer pH. Difficult to isolate without an alginate-only control well.

P1000 pipettes deliver ±2% accuracy at full volume (±20 µL). The protocol specifies this volume precisely and students are instructed to be accurate.

A 5% volume error (±50 µL) changes CO₂ buffering capacity by ~5%, translating to roughly ±0.01–0.015 ΔpH. Much smaller than the biological signal but potentially significant in respiration where condition differences are ~0.06 ΔpH.

A consistently underfilled tube will show a slightly larger ΔpH signal (less buffer to absorb CO₂). Could contribute to one group appearing to have higher respiration rates than others.

One group's ΔpH values consistently higher than all others across all conditions, in the absence of other explanations.

P200 pipettes deliver ±2–3% at 100 µL. Within-row duplicate wells are pipetted consecutively, so replicate precision is typically better than absolute accuracy.

Indicator volume affects starting absorbance. A 10% volume error (±10 µL) changes T0 absorbance by ~10%, which propagates proportionally into the ΔpH calculation.

Systematic over/under-dosing shifts all wells equally, preserving relative ΔpH differences. The duplicates (R1 vs R2) reveal within-condition pipetting inconsistency directly.

Large replicate differences (Rep Diff > 0.05) in the Analyser output. Also detectable if T0 absorbance is systematically higher or lower than expected for the working indicator concentration.

The calibration curve is prepared from the same stock solution as the working indicator. Sigma-Aldrich reagents are standardised with known purity. The curve is validated against Southern Biological pH standards.

Aged indicator stocks can shift pKa behaviour. The calibration was run on a different date to the student session. Temperature affects indicator absorbance and its sensitivity to pH change.

pH values are systematically offset, but ΔpH relative to T0 is less affected because T0 and T1.5 use the same indicator batch. More serious if calibration shape is non-linear in the working range.

T0 pH value is outside expected range (e.g., 8.2 instead of 7.4 due to aged buffer oxidation). All condition pH values shifted equally relative to historical data.

P200 pipette at 100 µL has ±2–3% systematic error and ±1% random error. The critical step is consistent — the same volume is transferred to both replicate wells from the same tube.

Since 100 µL sample is added to 100 µL indicator in the well (1:1 dilution), a 5% volume error translates directly to a 5% dilution error, affecting the final absorbance and pH by ~5%.

Proportional shift in apparent ΔpH. The replicate wells are transferred consecutively, so rep differences directly measure transfer consistency for each condition.

R1 vs R2 replicate difference. A large Rep Diff here reflects transfer rather than incubation variability (incubation variability would affect both wells equally).

Fresh pipette tips are used for each transfer. Wells are physically separated. The plates are fresh (not reused), eliminating residual contamination.

If a tip is accidentally reused, CO₂-rich buffer from one condition could be introduced to another. This is an execution error with immediate consequences, but the protocol makes it unlikely.

A reused tip would typically affect only the well loaded immediately after the error. Detectable as an outlier that differs substantially from its replicate.

An anomalously high or low absorbance in one well of a pair, giving a very large Rep Diff with no other explanation.

Transfer time is short (seconds). CO₂ equilibration with air is slow compared to transfer speed. The protocol specifies loading the plate immediately and reading quickly.

An open tube loses CO₂ to the atmosphere faster than a capped one. In the respiration experiment, 60°C tubes in particular — specified to be taped shut — must be opened for transfer. Any delay between removing the lid and sampling allows CO₂ outgassing.

Underestimates ΔpH in respiration (apparent signal is lower than actual). More CO₂ lost from higher-temperature tubes (60°C most affected), potentially narrowing the gap between conditions.

60°C ΔpH surprisingly close to 40°C values despite higher temperature. Slower loaders will show systematically smaller ΔpH signals than faster loaders for the same conditions.

LED sources produce minimal heat compared to incandescent lights. The boxes are open at the bottom for ventilation. A 1.5-hour incubation at room temperature is not expected to raise tube temperature significantly.

High-intensity LEDs (especially Blue at 214 mW/m²) can raise enclosed temperatures by 2–5°C over 90 minutes. Higher temperature slightly increases photosynthesis and also slightly increases dark respiration, partially offsetting each other.

A 3°C elevation increases enzymatic activity by roughly 10% (Q₁₀ ≈ 2). This adds ~0.03 ΔpH to high-intensity conditions — noticeable but not enough to change the ranking.

Difficult from pH data alone. Would require placing a thermometer inside the box for the full incubation period.

We have measured actual photon flux with a spectroradiometer (UPRtek MK350N). These values allow the intensity confound to be explicitly quantified rather than ignored.

Measured range: Blue 214 mW/m² → Yellow 35 mW/m² — a 5× difference. If Blue is best for photosynthesis but also the brightest, we cannot fully separate the wavelength and intensity effects without normalising.

The conclusion "Blue light is most effective for photosynthesis" is confounded — it might partly be because Blue is also the most intense. Wavelength and intensity effects are inseparable without intensity-controlled LEDs.

Compare actual intensities to ΔpH ranking. A condition with higher ΔpH than expected from its wavelength absorption might be explained by higher photon flux.

Blue LED at 462 nm overlaps well with chlorophyll b absorption (453 nm). The broad absorption peaks of chlorophyll mean wavelengths within ±20–30 nm of the peak are still absorbed efficiently.

Red LED peaks at 630 nm, which is 32 nm below chlorophyll a's red peak (662 nm). Absorption at 630 nm is approximately 70–80% of maximum, underestimating red light's true photosynthetic potential compared to a 660 nm source.

Red light ΔpH is likely an underestimate. If a true 660 nm LED was used, red might rank higher. This is a design limitation — it cannot be corrected through better technique.

Cannot be detected from pH data alone — requires prior knowledge of the chlorophyll absorption spectrum and LED characterisation data.

Water baths are set before the session and allowed to equilibrate. The tubes are submerged for the full incubation. Lids are taped to maintain heat.

Measured actual ranges: 60°C→57–59°C; 40°C→38–40°C; 20°C→21–22°C; 0°C→0–0.6°C. The 60°C bath shows the largest systematic offset (up to 3°C below setpoint). Shared baths mean multiple tubes compete for thermal equilibration when loaded simultaneously.

A 3°C offset at 60°C slightly underestimates enzyme denaturation at that temperature. However, at 57–59°C significant denaturation still occurs. The 40°C→60°C distinction is still biologically meaningful.

Measure actual bath temperature with a thermometer. Compare ΔpH at stated vs actual temperature — does it match the Q₁₀ predicted curve?

An ice-water mixture at equilibrium is thermodynamically locked at 0°C as long as ice remains — phase equilibrium is an unusually reliable temperature control compared to electronic thermostats.

Measured range: 0–0.6°C. The small deviation occurs where ice has melted locally around the tube. Temperature variation of 0.6°C has a negligible effect on yeast activity (Q₁₀ effect would be <1%).

Even a 2°C increase would change the 0°C reference ΔpH by only ~4% — barely detectable above noise. Ice baths are the most reliable temperature condition in this experiment.

Ice melting completely well before the end of incubation. The 0°C ΔpH would be larger than historical reference if incubation temperature rose significantly.

The Accuris MR9600 is calibrated against NIST-traceable absorbance standards. The instrument uses a blank auto-correction at the start of each run. Within-run CV is typically <1% for OD values in the working range.

Instrument drift between calibration sessions. Lamp aging shifts absolute absorbance values slightly. Filter bandpass variations affect readings when indicator absorbance spectra are broad.

Systematic offset applies equally to all wells — ΔpH (which uses T0 as reference from the same plate) is largely protected. Between-session comparisons are more vulnerable than within-session comparisons.

Empty wells (background) should read ~0.037–0.042 at 595 nm and ~0.040–0.043 at 450 nm. Values outside this range suggest instrument drift or dirty plate optics.

All tubes in a group are filled from the same buffer stock at the same time. T0 is measured at exactly that moment, before any incubation begins.

T0 represents the pH at loading time. If CO₂ equilibration is still occurring (buffer was freshly opened), the true starting pH of each tube may differ slightly from the T0 vial measured seconds later.

A T0 error shifts all ΔpH values equally up or down. Rankings of conditions are preserved. This is the single biggest issue seen in this class's data — G07's T0 abs of 0.048 (should be ~0.35) made all results uninterpretable.

T0 absorbance outside the expected range (0.30–0.60 for photo; 0.06–0.12 for resp at 450 nm). Very large or very small T0 rep diff. T0 pH far from expected buffer pH.

All groups use buffer from the same batch, prepared at the same time by the facilitator. ΔpH is calculated relative to each group's own T0, which partially corrects for between-group starting pH differences.

In this session, photosynthesis T0 absorbances ranged from 0.342 (pH~7.42) to 0.547 (pH~7.79). A T0 pH difference of 0.37 means groups are on different parts of the calibration curve, affecting sensitivity. Groups starting higher may show larger ΔpH simply because they have more room to move on the curve.

Groups that loaded different volumes, or used buffer from a slightly different time point, will have systematically different absolute pH values. Comparison of absolute ΔpH between groups requires this assumption to hold.

Compare T0 absorbance values across all groups. In the Feb 25 session: G02=0.361, G03=0.547, G04=0.381, G06=0.342 — G03 is a clear outlier.