Short answer: viable starter counts and acidification performance describe different parts of a yoghurt culture system; neither can be calculated reliably from the other.
This guide develops a focused reader question within the Yoghurt subject guide. It adds an explicit measurement chain, original worked examples, changed-context checks and limits. Read it with The Gold Standard of Education | Learning That Transfers into Life and the How X Works.
Every dataset and calculation below is an invented teaching case. The article does not certify a food, laboratory result, product, network or service. Use the current method, responsible specialist and applicable authority for a real decision.
eduKateSG · Yoghurt
Find your next learning step
Keep colony-count methods, acidification measures, strain identity, time and process conditions attached to every comparison.
ROUTE 1 · CHAPTERS 1–6
Define the measurements
Separate viability, acidification, strains, plates and logarithmic counts.
ROUTE 2 · CHAPTERS 7–11
Compare process evidence
Handle thresholds, acidity, inocula, cocultures and sampling times.
ROUTE 3 · CHAPTERS 12–17
Change the context
Test trajectories, activity, process stage, replication, limits and storage.
ROUTE 4 · CHAPTERS 18–22
Communicate and continue
Display evidence, limit claims, repair misconceptions and follow sources.
Chapter contents
Define the measurements · Chapters 1–6
1. Start with the short answer
2. Define viability operationally
3. Define acidification performance
4. Keep organism identity and strain visible
Compare process evidence · Chapters 7–11
7. Measure time to threshold without inventing precision
8. Separate pH from titratable acidity
9. Do not infer performance from the initial count alone
Change the context · Chapters 12–17
12. Changed context: same endpoint, different path
13. Changed context: same count, different activity
14. Changed context: counts after fermentation
15. Use replicates at the right level
16. Separate detection limits from zero
17. Keep storage observations from rewriting fermentation history
Communicate and continue · Chapters 18–22
18. Build a two-axis evidence display
19. Avoid food-safety and health leaps
20. Repair common misconceptions
A viable-cell count and an acidification record are related observations, but they are not interchangeable. Colony-forming units estimate cells able to form colonies under the declared culture method. Acidification records how the inoculated milk system changes over time under specified conditions. A sample can contain many culturable cells yet acidify slowly, or show rapid acidification even when a later plate count is lower.
Begin with a measurement ledger. Name the quantity, its unit, the sample or observation that produced it, the time window and the transformation rule. Mark each input as observed, stipulated, estimated or assumed. This stops a familiar label from smuggling in a new denominator or a stronger claim. A reader should be able to replace one input and see exactly which conclusion changes. Read the ledger as a sentence before calculating; if the sentence is incomplete, the arithmetic is premature.
In an invented comparison, culture A begins at 7.2 log10 CFU per gram and reaches pH 4.6 in 5.0 hours. Culture B begins at 6.8 log10 CFU per gram and reaches the same endpoint in 4.2 hours. The lower starting plate count does not prevent the faster observed acidification in this experiment.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to swap the endpoint for a fixed-time pH change and explain why the ranking may change. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
Viability is not a glow attached to every cell. A colony-count result depends on sampling, dilution, medium, incubation atmosphere, temperature, duration and the rule for countable plates. It estimates cells able to produce colonies under that procedure; it does not count every metabolically active, injured or nonculturable cell.
Keep the mechanism beside the number. Ask what physical, biological, statistical or behavioural relationship connects the observation to the conclusion, and which conditions could break that link. A correct calculation can answer the wrong question when a method selects only part of a sample or a changed setting alters the process. Draw one arrow for every inferential step, then challenge each arrow with a counterexample. State why the relationship is plausible and where it stops.
A 10^-6 dilution produces 84 colonies from a one-millilitre plated portion. Under the fictional method, the estimate is 84 million CFU per gram after applying the declared sample conversion. Calling it 84 million living cells would be stronger than the method supports.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to change the plated volume to 0.1 millilitres and reconstruct the conversion without changing the biological claim. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
Acidification can be expressed as time to a stated pH, change in pH over an interval, change in titratable acidity, or a fitted rate over a declared segment. These are different summaries. The endpoint and sampling schedule must travel with the number.
Use an independent check that follows a different route. Reconstruct totals from components, compare a bound with a point estimate, inspect a distribution behind an average, or solve a changed case with the original definitions. Agreement does not prove every assumption, but disagreement locates a definition, unit or boundary that needs repair. Write the expected check before seeing the answer so the check does not become a second performance of the same mistake.
Batch C moves from pH 6.55 to 5.20 in two hours, an average change of 0.675 pH units per hour. Because pH is logarithmic, this quotient is a descriptive slope in pH units, not a linear acid-production rate or a mass of lactic acid.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to repeat the summary with titratable-acidity data and keep the two measurement scales separate. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 4 OF 22 · Define the measurements
4. Keep organism identity and strain visible
Species names orient the comparison, while strain identity can matter to growth, cooperation and metabolite production. A result from one named pair of starter strains should not be generalized to every commercial culture carrying the same species names.
Match the claim to the evidence. A tidy decimal cannot repair selective observation, and one convenient run cannot establish every product, participant or network. Report the result, the conditions that make it interpretable, and the next observation needed for the actual decision. Separate the numerical conclusion from the practical conclusion. The first says what follows from the supplied model; the second names the authority, replication, context or threshold still required before action.
A paper compares strain S1 with strain S2 in one milk base. A classroom table that relabels both simply as yoghurt bacteria destroys the tested contrast. Retaining strain, medium and inoculum ratio preserves the experiment scope.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to replace one strain while holding the other inputs fixed and identify which earlier conclusion no longer transfers. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 5 OF 22 · Define the measurements
5. Worked example: count colonies before fermentation
Start with the raw plates, dilution and plated mass or volume. Choose a declared countable range, average only compatible plates and show the conversion. Do not use a convenient plate while hiding an overcrowded or empty companion plate.
Begin with a measurement ledger. Name the quantity, its unit, the sample or observation that produced it, the time window and the transformation rule. Mark each input as observed, stipulated, estimated or assumed. This stops a familiar label from smuggling in a new denominator or a stronger claim. A reader should be able to replace one input and see exactly which conclusion changes. Read the ledger as a sentence before calculating; if the sentence is incomplete, the arithmetic is premature.
Two compatible plates from the 10^-6 dilution contain 92 and 108 colonies from 1 mL portions. Their mean is 100, giving 1.00×10^8 CFU per mL in the stipulated direct-conversion model. The arithmetic is exact to the model; sampling and culture uncertainty remain.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to use counts of 18 and 240 under a method that accepts 30–300 and explain why automatic averaging is questionable. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 6 OF 22 · Define the measurements
6. Worked example: log changes need careful language
Microbial counts are often expressed on a base-ten logarithmic scale. Subtracting log values gives a log change; exponentiating gives the fold ratio. Neither alone explains why the population changed.
Keep the mechanism beside the number. Ask what physical, biological, statistical or behavioural relationship connects the observation to the conclusion, and which conditions could break that link. A correct calculation can answer the wrong question when a method selects only part of a sample or a changed setting alters the process. Draw one arrow for every inferential step, then challenge each arrow with a counterexample. State why the relationship is plausible and where it stops.
A count falls from 8.0 to 6.5 log10 CFU per gram. The change is -1.5 log units and the final count is about 10^-1.5, or 3.16%, of the initial count: roughly a 31.6-fold reduction. It is not a 1.5% decrease.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to calculate the fold change from 6.2 to 7.0 log10 CFU per gram and label direction correctly. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 7 OF 22 · Compare process evidence
7. Measure time to threshold without inventing precision
When samples are taken intermittently, the true threshold-crossing time lies between observations unless a justified model is fitted. Reporting the later observation as the exact crossing overstates what was seen.
Use an independent check that follows a different route. Reconstruct totals from components, compare a bound with a point estimate, inspect a distribution behind an average, or solve a changed case with the original definitions. Agreement does not prove every assumption, but disagreement locates a definition, unit or boundary that needs repair. Write the expected check before seeing the answer so the check does not become a second performance of the same mistake.
A batch is above pH 4.6 at 4.0 hours and below it at 4.5 hours. The observed crossing is bounded to that half-hour interval. Linear interpolation may supply an estimate, but it needs the two readings and an explicit local-linearity assumption.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to widen the sampling interval to one hour and show how the evidential bound changes. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
pH concerns hydrogen-ion activity; titratable acidity records the amount of titrant needed to reach a declared endpoint under a method. Buffering means equal pH values do not force equal titratable-acidity values.
Match the claim to the evidence. A tidy decimal cannot repair selective observation, and one convenient run cannot establish every product, participant or network. Report the result, the conditions that make it interpretable, and the next observation needed for the actual decision. Separate the numerical conclusion from the practical conclusion. The first says what follows from the supplied model; the second names the authority, replication, context or threshold still required before action.
Two fictional yoghurts both read pH 4.40. Sample D requires 8.2 mL of the stipulated titrant and sample E 10.1 mL for the same mass and endpoint. The results can coexist without contradiction.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to hold titratable acidity equal while changing pH and explain why one cannot be calculated uniquely from the other. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 9 OF 22 · Compare process evidence
9. Do not infer performance from the initial count alone
More inoculated culturable cells can shorten a lag or increase activity, but the relationship is not a universal proportional law. Strain physiology, inoculum history, temperature, milk composition and cooperation can alter the response.
Begin with a measurement ledger. Name the quantity, its unit, the sample or observation that produced it, the time window and the transformation rule. Mark each input as observed, stipulated, estimated or assumed. This stops a familiar label from smuggling in a new denominator or a stronger claim. A reader should be able to replace one input and see exactly which conclusion changes. Read the ledger as a sentence before calculating; if the sentence is incomplete, the arithmetic is premature.
Culture F starts with twice the CFU of culture G, yet reaches the endpoint only 10% sooner. Dividing time by initial count would imply a false inverse law. The observation supports this pair under these conditions only.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to change temperature for one culture and identify the confounding introduced into the count-performance comparison. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 10 OF 22 · Compare process evidence
10. Keep single cultures and cocultures distinct
A coculture is a biological system with possible interaction. Its acidification curve is not necessarily the arithmetic sum of two monoculture curves. Shared metabolites and changing conditions can produce non-additive behaviour.
Keep the mechanism beside the number. Ask what physical, biological, statistical or behavioural relationship connects the observation to the conclusion, and which conditions could break that link. A correct calculation can answer the wrong question when a method selects only part of a sample or a changed setting alters the process. Draw one arrow for every inferential step, then challenge each arrow with a counterexample. State why the relationship is plausible and where it stops.
At hour three, monocultures lower pH by 0.40 and 0.55 units. The coculture lowers it by 1.20 units. Adding 0.95 describes a naive reference, not a physical law; the 0.25 difference prompts an interaction hypothesis rather than proving a mechanism.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to change inoculum ratio while keeping total inoculum constant and specify the comparison needed. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 11 OF 22 · Compare process evidence
11. Align sampling times before comparing curves
Curves sampled at different times cannot be compared by selecting whichever readings look closest. Use common observation times, a justified interpolation method, or interval statements. Preserve any phase where one series is missing.
Use an independent check that follows a different route. Reconstruct totals from components, compare a bound with a point estimate, inspect a distribution behind an average, or solve a changed case with the original definitions. Agreement does not prove every assumption, but disagreement locates a definition, unit or boundary that needs repair. Write the expected check before seeing the answer so the check does not become a second performance of the same mistake.
Batch H is observed hourly; batch I every ninety minutes. At three hours both have direct readings, so that point is comparable. A claimed two-hour difference that uses an interpolated I value needs that label and model.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to remove the common three-hour point and choose a transparent comparison rule. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 12 OF 22 · Change the context
12. Changed context: same endpoint, different path
Equal final pH does not mean equal kinetics. Lag, steepest change, time near an intermediate value and post-endpoint behaviour can differ. A process question often needs the whole trajectory.
Match the claim to the evidence. A tidy decimal cannot repair selective observation, and one convenient run cannot establish every product, participant or network. Report the result, the conditions that make it interpretable, and the next observation needed for the actual decision. Separate the numerical conclusion from the practical conclusion. The first says what follows from the supplied model; the second names the authority, replication, context or threshold still required before action.
Batches J and K both reach pH 4.5 at six hours. J changes steadily; K remains near 6.3 for three hours and then falls quickly. One endpoint hides the different histories.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to compare area under a declared response curve or time in a specified pH band without calling it cell count. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 13 OF 22 · Change the context
13. Changed context: same count, different activity
Equal colony counts under one plating method do not establish equal metabolic state in milk. Cells can differ in physiological history, stress and interaction with the medium.
Begin with a measurement ledger. Name the quantity, its unit, the sample or observation that produced it, the time window and the transformation rule. Mark each input as observed, stipulated, estimated or assumed. This stops a familiar label from smuggling in a new denominator or a stronger claim. A reader should be able to replace one input and see exactly which conclusion changes. Read the ledger as a sentence before calculating; if the sentence is incomplete, the arithmetic is premature.
Two inocula each measure 7.0 log10 CFU per gram. One reaches pH 5.0 in three hours and the other in four and a half. The result shows that equal operational counts did not equalize the observed acidification performance.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to repeat the comparison after standardized preconditioning and explain what that controls. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 14 OF 22 · Change the context
14. Changed context: counts after fermentation
A count measured after fermentation answers a different question from the starting inoculum count. Acid conditions and sampling position can influence recovery. Pair every count with its process time.
Keep the mechanism beside the number. Ask what physical, biological, statistical or behavioural relationship connects the observation to the conclusion, and which conditions could break that link. A correct calculation can answer the wrong question when a method selects only part of a sample or a changed setting alters the process. Draw one arrow for every inferential step, then challenge each arrow with a counterexample. State why the relationship is plausible and where it stops.
A batch begins at 6.9 log10 CFU per gram and ends at 8.1. Another begins at 7.3 and ends at 7.8. The second starts higher but the first ends higher; rankings depend on the stated time.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to calculate net log changes and resist interpreting them as unique growth rates without time-course data. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
Multiple plates from one dilution are technical repeats. Independent fermentation vessels, culture preparations or product units address broader variability. Treating ten plates from one jar as ten production batches overstates independence.
Use an independent check that follows a different route. Reconstruct totals from components, compare a bound with a point estimate, inspect a distribution behind an average, or solve a changed case with the original definitions. Agreement does not prove every assumption, but disagreement locates a definition, unit or boundary that needs repair. Write the expected check before seeing the answer so the check does not become a second performance of the same mistake.
Three vessels each provide duplicate plates. The analysis should preserve vessel means and between-vessel variation. Pooling all six plate counts without vessel identity hides the experimental structure.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to compare technical and biological replication in a new nested dataset. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
No colonies observed does not prove the sample contains no viable cells. The plated amount and dilution set a detection boundary. Report the less-than statement that follows from the method.
Match the claim to the evidence. A tidy decimal cannot repair selective observation, and one convenient run cannot establish every product, participant or network. Report the result, the conditions that make it interpretable, and the next observation needed for the actual decision. Separate the numerical conclusion from the practical conclusion. The first says what follows from the supplied model; the second names the authority, replication, context or threshold still required before action.
If 0.1 g equivalent is plated and no colony appears, the direct one-colony boundary is 10 CFU per gram. Reporting zero CFU per gram converts non-detection into an impossible exact absence claim.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to increase the plated equivalent to one gram and show the changed boundary. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 17 OF 22 · Change the context
17. Keep storage observations from rewriting fermentation history
A stored product later count and acidity reflect what happened after fermentation as well as the starting process. Temperature, time, packaging and sample handling need to remain beside the result.
Begin with a measurement ledger. Name the quantity, its unit, the sample or observation that produced it, the time window and the transformation rule. Mark each input as observed, stipulated, estimated or assumed. This stops a familiar label from smuggling in a new denominator or a stronger claim. A reader should be able to replace one input and see exactly which conclusion changes. Read the ledger as a sentence before calculating; if the sentence is incomplete, the arithmetic is premature.
At day 1 a product has 8.2 log10 CFU per gram and pH 4.45; at day 21 it has 7.1 and pH 4.25. The paired change is descriptive and does not by itself identify the mechanism or establish safety.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to compare two storage temperatures without treating an uncontrolled shelf record as a production recommendation. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
Plot acidification against time on one axis and viable count at matched sampling points on another panel. Do not place unlike units on one unlabeled axis or imply that visual overlap proves causation.
Keep the mechanism beside the number. Ask what physical, biological, statistical or behavioural relationship connects the observation to the conclusion, and which conditions could break that link. A correct calculation can answer the wrong question when a method selects only part of a sample or a changed setting alters the process. Draw one arrow for every inferential step, then challenge each arrow with a counterexample. State why the relationship is plausible and where it stops.
At 0, 2, 4 and 6 hours, a fictional study records both pH and log CFU. The paired table lets readers see whether faster pH change coincides with count growth while preserving two measurement scales.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to shift the count samples by thirty minutes and label the mismatch before interpreting it. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
Starter viability and acidification are quality and process evidence under defined methods. They do not by themselves certify pathogen absence, probiotic benefit, shelf life or suitability for an individual.
Use an independent check that follows a different route. Reconstruct totals from components, compare a bound with a point estimate, inspect a distribution behind an average, or solve a changed case with the original definitions. Agreement does not prove every assumption, but disagreement locates a definition, unit or boundary that needs repair. Write the expected check before seeing the answer so the check does not become a second performance of the same mistake.
A batch reaches its acidification endpoint quickly and has a high starter count. Those facts do not replace required hazard controls, composition evidence, regulatory checks or clinical evidence.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to rewrite a promotional claim so it reports only the measured starter and acidification outcomes. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
A CFU is not one guaranteed cell; a log difference is not a percentage-point difference; pH change is not acid mass; and endpoint time is not viability. Each shortcut removes a method or denominator.
Match the claim to the evidence. A tidy decimal cannot repair selective observation, and one convenient run cannot establish every product, participant or network. Report the result, the conditions that make it interpretable, and the next observation needed for the actual decision. Separate the numerical conclusion from the practical conclusion. The first says what follows from the supplied model; the second names the authority, replication, context or threshold still required before action.
Ask a learner to diagnose: Batch L has twice the CFU, so it must make acid twice as fast. The first repair is to separate the two measured variables; the second is to test the relationship rather than assume it.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to construct a counterexample with equal counts and unequal endpoint times. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
CHAPTER 21 OF 22 · Communicate and continue
21. Connect the existing yoghurt branches
The water-retention guide separates pellet fraction, retained water and solids. The root develops pH, titratable acidity and culture partnerships. This branch adds the missing operational distinction between culturable count and acidification performance.
Begin with a measurement ledger. Name the quantity, its unit, the sample or observation that produced it, the time window and the transformation rule. Mark each input as observed, stipulated, estimated or assumed. This stops a familiar label from smuggling in a new denominator or a stronger claim. A reader should be able to replace one input and see exactly which conclusion changes. Read the ledger as a sentence before calculating; if the sentence is incomplete, the arithmetic is premature.
A complete study can contain all three kinds of evidence while keeping their methods separate. Stronger explanation comes from connection, not substitution.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to choose which branch answers a question about cells, acidification or separated liquid. State what would change the result and what the example cannot establish. That limit belongs inside the answer.
ISO 7889:2026 specifies colony-count enumeration of the characteristic yoghurt microorganisms. The Codex Standard for Fermented Milks supplies product context, and the root links primary studies of named strains and cocultures. Sources were checked on 26 October 2026.
Keep the mechanism beside the number. Ask what physical, biological, statistical or behavioural relationship connects the observation to the conclusion, and which conditions could break that link. A correct calculation can answer the wrong question when a method selects only part of a sample or a changed setting alters the process. Draw one arrow for every inferential step, then challenge each arrow with a counterexample. State why the relationship is plausible and where it stops.
The teaching calculations do not reproduce a laboratory protocol or declare compliance. Method validation, competent microbiology and the current applicable standard govern real enumeration.
Before moving on, record the observation that would distinguish the two most plausible explanations. If no feasible observation exists, narrow the claim instead of pretending the ambiguity has disappeared.
A changed-context transfer check is to audit an unfamiliar paper by extracting strain, medium, initial count, acidification metric, sampling times and claim limit. State what would change the result and what the example cannot establish. That limit belongs inside the answer.