The Dish-Soap Antibacterial Fair Test
Four groups test whether three dish-washing liquids inhibit bacterial growth. Each group pours a lawn of the same bacterium on a nutrient agar plate, places four paper discs on it soaked in liquid A, liquid B, liquid C and sterile water, and incubates at 30 °C for 24 h. Each disc receives 10 µL. Zone of inhibition diameters in mm:
A: 14, 15, 13 B: 9, 10, 10 C: 16, 4, 15 water: 0, 0, 0
Liquid A is used undiluted, as sold. Liquid B is labelled 'concentrated - dilute 1:10 before use' but was also applied undiluted. Liquid C came from a bottle opened a year earlier. One plate in the C series shows colonies of a different, white bacterium spreading from the edge. The class concludes that C is the strongest antibacterial agent and A the second strongest, and writes this up as their result.
1. Identify the two flaws that most seriously undermine the ranking of A, B and C, and state the direction in which each flaw biases the comparison.
Read the worked answer
First, the concentrations are not comparable, and two of them are not even known. B is meant to be diluted tenfold but was applied neat, so its zone reflects a tenfold overdose of active ingredient. C came from a year-old opened bottle, so its surfactant may have degraded or evaporated and its true strength is unknown. Zone size depends on how much agent is applied, so a small zone for B and a large zone for C say nothing about intrinsic potency. Direction of bias: B is unfairly penalised, and C is tested under an unknown and probably reduced activity - yet C still produced the largest zones, so the ranking C > A is confounded rather than supported.
Second, the four discs share one plate instead of being independent replicates. If the discs are closer than the sum of their zone radii, the zones merge and each is truncated, and each treatment's value is influenced by its neighbours' diffusion. A single plate also means one lawn, one batch of medium and one incubation position, so any plate-level accident contaminates all four treatments on that plate at once. Direction: unpredictable, but it inflates apparent variability and can manufacture or hide differences. Replication is three plates, which is thin for any ranking.
2. What should be done about (i) the 4 mm reading for C together with the contaminated plate, and (ii) the concentration problem? Give a specific revised design.
Read the worked answer
(i) The contaminated plate is a failed run: the white colonies show the culture or plate was compromised, so it is discarded for a stated cause, not because its number is inconvenient. That leaves C with only two valid replicates (16 and 15 mm). The 4 mm value should be attached to the failed plate and investigated - misplaced disc, no inoculum there, or a zone truncated by the plate edge - and reported as such, never silently averaged in or deleted as an outlier.
(ii) Compare like with like: dilute all three liquids to the same active-ingredient concentration (for example 1 % v/v in sterile water), and test each at several dilutions (1 %, 0.5 %, 0.1 %) so that a dose-response exists and a single concentration cannot decide the ranking.
Revised design: five independent plates per treatment; one central disc per plate to remove merging and position effects; identical 10 µL volume on every disc; standardised inoculum density; a water control and a reference antibiotic disc; randomised incubation positions; 30 °C for 24 h; the person measuring blind to which liquid is which. Report mean zone with the range, and repeat the whole experiment on a second day to show the result is reproducible.
3. The teacher says that even with perfect zones, zone width is not the same as killing power. Explain two mechanisms that could make a potent agent give a small zone, and describe a measurement that would test killing directly.
Read the worked answer
A zone marks the distance an agent travels before its concentration falls below the inhibitory threshold, so it mixes potency with mobility.
- Molecular size and viscosity: a large or strongly surface-active molecule diffuses slowly through agar, so it stays concentrated near the disc and produces a narrow zone even where the local concentration is lethal.
- Binding and sequestration: agents that stick to agar polysaccharides or to proteins in the medium are removed from free solution, again shrinking the zone without any change in intrinsic activity.
- A third confusion: a clear zone shows inhibition (bacteriostasis), not death. Cells inside the zone may be alive but not dividing.
To test killing, sample the survivors. Take a loop from inside the clear zone, wash it, dilute it in sterile saline and plate known volumes on agar containing no agent; count colonies and compare with the untreated control to get a fractional survival. Alternatively run a liquid minimum inhibitory concentration series followed by a minimum bactericidal concentration series with a fixed inoculum and contact time. These give numbers comparable between agents regardless of how fast each one diffuses.