CFU counting
Ten-fold dilutions of the sample are made in sterile diluent and 100 µL of each of three consecutive dilutions is spread on its own plate, usually in duplicate. Only plates with 30–300 colonies are counted and the titre is the weighted mean of those plates.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
Spread-plate counts
Ten-fold dilutions of the sample are made in sterile diluent and 100 µL of each of three consecutive dilutions is spread on its own plate, usually in duplicate. Only plates with 30–300 colonies are counted and the titre is the weighted mean of those plates.
How it differs from the standard method
- One plate per dilution per replicate: more plates and media, but the largest counts and the smallest Poisson error.
- Colonies are on the surface and can be picked afterwards.
- A pour plate (1 mL mixed into molten agar) is the same calculation with a ten-fold larger volume and embedded colonies.
- Counted after a full overnight, 16–24 h for E. coli.
1Counting what can grow
A viable count answers a question that optical density and a counting chamber cannot: how many cells in this sample are alive and able to divide. It does so indirectly.
2Why only 30 to 300
Colonies are independent rare events, so the number on a plate follows Poisson statistics and its relative error is about 1/√N. A plate with 9 colonies is uncertain by a third before any pipetting error; one with 100 by a tenth.
Too few
Poisson error above ~18%, and one stray colony shifts the answer. Report it only if nothing better exists, and say so.
Countable
Precise enough to quote and sparse enough to count every colony. Average replicates and use all plates in range.
Too numerous
Colonies merge and the count runs low. A crowded plate is tempting because the number is big, but it is biased.
3How a dilution series goes wrong
The ten-fold steps are the weak point, because an error in one tube is inherited by every tube after it. Two mistakes dominate.
4Spread plates and drop plates
Spread plating puts 100 µL of each of a few dilutions on its own plate. Drop plating (the Miles and Misra method and its descendants) puts 5–20 µL drops of every dilution on a single plate, usually in replicate rows.
- TNTC
- Too numerous to count. Recorded, not estimated, for any plate above the countable range.
- Weighted mean
- When two dilutions both fall in range, the total colonies on all counted plates divided by the total volume-times-dilution they represent. It weights each plate by how many colonies it contributes.
- Plating efficiency
- The fraction of viable cells that actually form a colony under the conditions used. Stressed, injured or fastidious cells can be alive and still not form one.
Key terms
Practise these as flashcards →A cell, chain or clump that gives rise to one colony. CFU/mL is the viable count of a suspension.
How much the tube plated has been diluted from the original: 10⁻⁵ means one part in 100,000. Multiply back by its inverse.
The colony numbers per plate (about 30–300 on 90 mm) or per spot (about 3–30 for 10 µL) where counts are both precise and complete.
The scatter of a count of independent events: a relative standard error of about 1/√N, where N is the total number of colonies counted.
Liquid from a denser tube carried into the next on a reused tip or pipette, which makes each dilution step smaller than intended.
Counting from small drops of each dilution spotted onto one plate, rather than one plate per dilution.
When to use it
- You need the number of living, growing cells, not total biomass: survival after a treatment, a killing curve, a challenge dose, or the titre of a stock.
- Calibrating optical density for a strain, by plating a culture of known OD600 and counting the CFU behind it.
- Measuring frequencies, such as mutants, transformants or plasmid loss, by plating the same culture on selective and non-selective media.
- Many samples or a long time course where one plate per dilution is impractical: drop plating.
Compared with alternatives
Uses the same spreading and pouring, but for isolating single colonies rather than counting them.
Optical density follows growth continuously and instantly, but counts dead cells and debris too and saturates at high density. CFU counts calibrate it.
A counting chamber or automated counter gives total (and dye-excluding) cells within minutes, without waiting for growth; it does not show whether cells can divide.
The dilution series is where most CFU errors are made: mixing, tip changes and the arithmetic of ten-fold steps.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
Ten-fold dilutions of the sample are made in sterile diluent and 100 µL of each of three consecutive dilutions is spread on its own plate, usually in duplicate. Only plates with 30–300 colonies are counted and the titre is the weighted mean of those plates.
- One plate per dilution per replicate: more plates and media, but the largest counts and the smallest Poisson error.
- Colonies are on the surface and can be picked afterwards.
- A pour plate (1 mL mixed into molten agar) is the same calculation with a ten-fold larger volume and embedded colonies.
- Counted after a full overnight, 16–24 h for E. coli.