Aerobic Plate Count (CFU) Calculator
Enter the two plated dilution levels, replicate counts, TNTC plates, inoculum volume, and sample equivalent. The calculator estimates APC as CFU/mL or CFU/g and shows likelihood-based confidence limits.
Estimated aerobic plate count19,181.82
- Log10 aerobic plate count
- 4.28
- 95% confidence interval - lower limit
- 17,409.53
- 95% confidence interval - upper limit
- 21,070.49
- Original-sample amount on exact-count plates
- 0.022 original mL or g
- Reportable result
- 19,182 CFU/mL; 95% confidence interval 17,410 CFU/mL to 21,070 CFU/mL.
How to use this calculator
- Choose CFU/mL for liquids or CFU/g for solids, then enter the original sample equivalent in each milliliter of the initial suspension.
- Enter the first plated dilution exponent, dilution step, and inoculum volume used for each plate.
- Enter exact-count plate numbers and their total colonies at each dilution, including plates with zero colonies.
- Enter TNTC plate counts separately instead of assigning them an invented colony count.
- Review the estimate, confidence limits, reportable wording, and any dilution-consistency notes.
How the aerobic plate count is calculated
The calculation starts by converting each plated dilution into the amount of original sample represented on one plate. If q is the original sample equivalent per milliliter of the initial suspension, v is the inoculum volume, e is the first plated dilution exponent, and s is the exponent step to the second dilution, then:
V1 = q x v x 10^e
V2 = q x v x 10^(e - s)
For ordinary exact-count plates with no TNTC observations, the weighted estimate is:
APC = (C1 + C2) / (n1 x V1 + n2 x V2)
Here C1 and C2 are the total colonies counted at the two dilution levels, and n1 and n2 are the numbers of exact-count plates. This is a weighted count, not a simple average of per-plate CFU results, because plates representing more original sample carry more exposure.
How TNTC plates are handled
A TNTC plate says the colony count was at least the laboratory's lower TNTC threshold. This calculator treats that as a censored Poisson observation rather than replacing it with 300 or another fixed number. It maximizes the log-likelihood:
l(mu) = C1 ln(mu V1) - n1 mu V1 + C2 ln(mu V2) - n2 mu V2 + t1 ln(Poisson(mu V1) >= L) + t2 ln(Poisson(mu V2) >= L)
The unknown mu is the original concentration in CFU/mL or CFU/g. The confidence limits are profile-likelihood limits: the calculator finds the lower and upper concentrations where 2 x [l(mu-hat) - l(mu)] = 3.841, the 95% cutoff for one parameter.
What moves the result most
The largest drivers are the dilution exponent, inoculum volume, sample equivalent, and the total original-sample exposure on exact-count plates. Inoculating 0.1 mL instead of 1 mL adds a tenfold factor. For a 1:10 food homogenate, entering 0.1 g/mL makes the result CFU/g in the original food rather than CFU/mL in the homogenate.
What this calculator leaves out
The math assumes independent Poisson colony formation after the stated dilutions and volumes. It does not correct for pipetting error, uneven mixing, injured organisms, selective media, colony spreading, clumps that form one colony, incubation conditions, contamination, or analyst counting rules. Laboratory SOPs and the applicable standard decide which plates are reportable and what TNTC threshold applies.
Worked example
Suppose an undiluted liquid sample is plated at 10^-2 and 10^-3 with 1 mL per plate. Two exact plates at 10^-2 have 383 total colonies, and two exact plates at 10^-3 have 39 total colonies. There are no TNTC plates.
The original-sample amount per plate is 0.01 mL at 10^-2 and 0.001 mL at 10^-3. The total exact-count exposure is 2 x 0.01 + 2 x 0.001 = 0.022 mL. The weighted estimate is (383 + 39) / 0.022 = 19,181.8 CFU/mL, reported here as about 19,182 CFU/mL. The likelihood-ratio 95% interval is about 17,410 to 21,070 CFU/mL, and log10 APC is 4.28.
Common questions
How do I calculate CFU/mL from two serial dilutions?
Convert each plate to the amount of original sample it represents, add the colonies across exact-count plates, and divide by the total original-sample amount on those plates. For adjacent tenfold dilutions, the less dilute plates usually carry ten times the exposure of the more dilute plates.
Should replicate plate counts be averaged or weighted?
Replicates at the same dilution can be added because they have the same exposure. When two dilutions are combined, a weighted calculation is preferred because each dilution represents a different amount of original sample. A simple average of per-plate CFU values can give the wrong weight to the more dilute plates.
What should I do when one or more plates are TNTC?
Enter TNTC plates in the TNTC fields and leave their colonies out of the exact-count totals. The calculator treats TNTC as a lower-bound observation using the threshold you enter. Do not replace a TNTC plate with 300 unless your method specifically instructs you to do that.
What is the difference between CFU/mL and CFU/g?
CFU/mL reports concentration in an original liquid sample. CFU/g reports concentration in the original solid sample, usually after a weighed portion is diluted into a homogenate. Enter the original sample equivalent in the homogenate, such as 0.1 g/mL for a 1:10 preparation, to keep the unit tied to the original food or material.
Why does inoculating 0.1 mL add another tenfold factor?
A plate that receives 0.1 mL receives one tenth as much suspension as a plate that receives 1 mL. For the same dilution exponent and same colony count, that smaller volume implies a concentration ten times higher. The calculator includes this through the inoculum volume used for each dilution.
Why does the calculator give a confidence interval?
Colony counts are discrete observations from a small plated portion of the original sample. Two plates with the same true concentration will not give exactly the same colonies every time. The likelihood interval shows the range of concentrations that remain reasonably compatible with the observed counts under the Poisson model.