Plate Colony Count and CFU/mL Calculator (Countable Range / ISO Weighted Average / Poisson Error)
Converting a colony count to CFU/mL takes a single calculation:
CFU/mL = colony count × dilution factor ÷ plating volume (mL)
Plating 100 µL means dividing by 0.1 — i.e., multiplying by 10 — so missing this step shifts the result by a full order of magnitude, yet the number still looks plausible. But there are three things worth explaining beyond that one step.
1. Countable Range: Statistics, Not Arbitrary Rules
The common guidance is “only count plates with 25–250 or 30–300 colonies.” This is not an arbitrary rule — it follows from the Poisson distribution:
Count n colonies, and the relative standard deviation is 1/√n.
| Colonies counted | Relative SD |
|---|---|
| 10 | ±32% |
| 30 | ±18% |
| 100 | ±10% |
| 250 | ±6.3% |
| 1000 | ±3.2% |
That is where the lower limit comes from — count 10 colonies and random fluctuation alone gives ±32% error; no matter how accurate your pipetting, you cannot recover from it, because this error comes from the counting itself, not from your technique.
The upper limit has a different cause: dense colonies crowd and merge, so what looks like one colony is actually several — a systematic undercount that gets worse as the plate gets more crowded. Estimating a number from a confluent plate is the worst possible approach: it simultaneously violates the linearity range and treats a guess as a measurement. The correct response is to replate at a higher dilution.
2. When Multiple Dilutions Are Countable, Use the Weighted Average
If plates at both 10⁻² and 10⁻³ fall in the countable range, don’t just pick one — both carry information, and discarding one wastes half your data. The ISO 4833-1 formula is:
N = ΣC ÷ [ V × (n₁ + 0.1 × n₂) × d ]
ΣC: sum of colonies on all plates included in the calculationn₁: number of plates at the lowest accepted dilutionn₂: number of plates at the next dilution (one order of magnitude higher)d: the dilution factor at the lowest accepted dilution (as a fraction; a 100-fold dilution gives d = 0.01) — note that this tool’s input uses the fold value (100); internally it converts to d by taking the reciprocal. Swapping the two gives a 100-fold error.V: plating volume per plate (mL)
The 0.1 is not an empirical coefficient — it is simply the ratio of dilution factors. A plate at the higher dilution represents one-tenth the original volume per colony, so it contributes only 0.1 of a plate’s weight in the average.
This tool automatically identifies countable plates, applies this formula, and lists which plates were included.
3. CFU Is Not Cell Count
A colony-forming unit (CFU) counts the units that actually produce a colony. Streptococcal chains, staphylococcal clusters, two cells stuck together — each gives only one colony. As a result, CFU systematically underestimates true cell number by an amount that depends on how much the strain clumps and how well vortexing disperses it.
This is one reason plate counts and OD₆₀₀ estimates often disagree: OD measures all light-scattering particles (live and dead alike), while CFU counts only viable cells capable of forming isolated colonies. The two track each other in log phase and diverge increasingly into stationary phase. Report which method you used.
References
The weighted-average formula is from ISO 4833-1:2013 (Microbiology of the food chain — pour plate method at 30 °C). The two common countable-range limits (25–250 and 30–300) come from that standard and from FDA BAM Chapter 3 (aerobic plate count), respectively. The two standards genuinely disagree; this tool makes the limits editable and defaults to 25–250. Use whichever standard your field follows.
The Poisson relative SD of 1/√n is a property of the distribution itself, independent of any standard.
Related Tools
For competent cell transformation efficiency (which also requires a plating-volume correction), see Transformation efficiency calculator; for estimating culture density from OD₆₀₀, see OD₆₀₀ to cell density; for preparing serial dilutions, see Dilution calculator; for doubling time, see Doubling time calculator.
FAQ
Why can only plates with 25–250 (or 30–300) colonies be counted?
The lower limit is set by statistics: count n colonies and the Poisson relative SD is 1/√n — 10 colonies gives ±32%, 30 gives ±18%. This error **comes from the counting itself, not from your technique**; even perfect pipetting cannot fix it. The upper limit has a different cause: dense colonies crowd and merge, so one apparent colony may actually be several — a systematic undercount that worsens with density. The two limits (25–250 and 30–300) come from different standards; this tool makes them editable.
The plate is confluent and uncountable — can I estimate a number?
No. An estimate is not a measurement, and it will always be biased low — once colonies are dense enough to merge, the count is already outside the linear range. The correct response is to replate at a higher dilution. If the sample cannot be repeated, report the result as "greater than [value]" rather than a specific number.
Plates at two different dilutions are both in the countable range — which one should I use?
Use both. ISO 4833-1 specifies a weighted average: N = ΣC ÷ [V × (n₁ + 0.1 × n₂) × d], where n₁ is the number of plates at the lowest accepted dilution, n₂ is the number at the next dilution, and d is the lowest accepted dilution factor. The 0.1 is not an empirical coefficient — it is the ratio of dilution factors: a plate at the higher dilution represents one-tenth the original volume per colony, so it contributes only 0.1 of a plate's weight. Picking one arbitrarily discards half your data.
Does plating 100 µL versus 1 mL change the calculation?
The formula is the same, but the divisor differs: CFU/mL = colony count ÷ (dilution factor × plating volume in mL). Plating 100 µL means dividing by 0.1 — equivalent to **multiplying by 10**. Missing this step shifts the result by a full order of magnitude; the number still looks reasonable in isolation, and the error only becomes apparent when comparing across experiments. This tool explicitly shows the correction factor whenever the plating volume is not 1 mL.
Are CFU and cell count the same thing?
No, and the direction of the discrepancy is always the same. A colony-forming unit counts "units that produced a colony" — streptococcal chains, staphylococcal clusters, two cells stuck together each give only one colony. So **CFU systematically underestimates true cell number** by an amount that depends on how much the strain clumps and how effectively vortexing disperses it. This is also why plate counts and OD₆₀₀ estimates often disagree: OD counts dead cells and debris as well.
How can I reduce counting error?
The only way is to **count more colonies** — either plate more replicates at each dilution, or choose the dilution that gives a higher count (still within the countable range). The relative SD is 1/√ΣC, where ΣC is the total colony count across all plates in the calculation, so three plates of 200 colonies each (ΣC = 600, ±4.1%) is far better than one. Improving pipette calibration does nothing for this error — it is not a pipetting error.
Should I enter the dilution factor as 100 or 1e-2?
This tool's "dilution factor" field takes the **fold value** — enter 100 to indicate a 100-fold dilution (10⁻²). Scientific notation such as 1e2 is also accepted. Using the fold value matches how you think during the experiment ("take 1 part and add 99 parts = 100-fold dilution") without requiring mental conversion to exponent form at data entry.
Related tools
Centrifuge k-Factor Calculator: Convert Run Time When Switching Rotors
Compute rotor k-factor from r_min, r_max, and rated speed; scale run time when switching rotors. RCF measures force — k accounts for path length too.
Animal Dose Calculator — mg/kg to Injection Volume & Batch Preparation
Enter dose (mg/kg), body weight, and stock concentration to get injection volume; supports batch preparation and shows maximum volumes by route.
Bacterial / Cell Doubling Time and Growth Rate Calculator
Enter two log-phase OD or cell count measurements to calculate doubling time, specific growth rate μ, number of generations, and predicted cell density.
Buy me a coffee