Why a Plate Reader's OD600 Reads Low
5 min read · Updated October 30, 2026
OD600 is not a cell count. It is how much light a sample scatters, and two properties of the measurement turn it into the wrong number without anything on the instrument saying so: the path the light takes through the sample, and the density above which scattering stops being proportional to it.
Both are arithmetic, and both are worth doing before any growth curve is compared with anybody else's.
The path length, in one worked number
Absorbance is proportional to path length. A cuvette is 1 cm by definition, and that is the path every published cells-per-OD figure assumes. A plate reader measures straight down through the liquid column, so the path is whatever the well is filled to — about 0.58 cm for 200 µL in a flat-bottomed 96-well plate.
So a reading of 0.42 taken in that plate is 0.42 / 0.58 = 0.724 in cuvette terms. At the usual E. coli convention of 8 × 10⁸ cells/mL per OD unit, the same sample is 3.36 × 10⁸ cells/mL if you treat the reading as a cuvette's, and 5.79 × 10⁸ cells/mL once the path is corrected. That is a 1.72-fold difference, and it is entirely an artefact of where the reading was taken.
Nothing about the two numbers looks wrong on its own, which is why the correction has to be applied by habit rather than when something seems off.
Above about 0.8, a dense culture reads low
Beer-Lambert behaviour holds while most photons are scattered at most once. Past roughly OD 0.8 in a 1 cm path, multiple scattering sends some of that light back into the detector, so the reading rises more slowly than the cell density does — and the culture reads lower than it is.
The consequence is systematic: every count derived from a reading above the linear range is an underestimate, and every growth curve flattens earlier than the culture did. The fix is to dilute into the linear range, read, and multiply back by the dilution factor, rather than to read the neat culture and trust it.
A reading of 1.2 is already past this, and so is anything taken off an overnight culture.
Cells per OD is a convention, not a constant
8 × 10⁸ cells/mL per OD for E. coli is a widely used round number, and the commonly quoted range is 5 × 10⁸ to 1 × 10⁹. It varies with strain, with growth phase — cells in stationary phase are smaller and scatter differently from cells in log phase — and with the instrument's optics.
That is a two-fold spread, so a cell count from OD is a good estimate and a poor measurement. When the number matters, calibrate your own factor by plating for colony-forming units at a known OD on your own instrument, and use that factor afterwards.
The back-dilution, worked
An overnight at OD 2.4, diluted into 50 mL of fresh medium to start at OD 0.05, needs 1.042 mL of culture and 48.96 mL of medium — a 48-fold dilution.
Two things make that number wrong in practice. The first is the one above: the overnight was read above the linear range, so it is denser than 2.4 and the inoculum is an underestimate. The second is that OD measures scattering particles rather than viable cells, and an overnight contains debris and dead cells that a fresh log-phase culture does not.
Both push the same way, which is why a back-dilution from an overnight reliably takes slightly longer to reach a target OD than the arithmetic says.
Frequently asked questions
How many cells per mL is an OD600 of 1.0, and how precise is it?
About 8 × 10⁸ cells/mL by the usual convention, with a commonly quoted range of 5 × 10⁸ to 1 × 10⁹. The figure assumes a 1 cm path and depends on strain, growth phase and instrument, so treat it as an estimate good to about two-fold unless you have calibrated against plate counts.
Why is the same culture lower in a 96-well plate than in a cuvette?
Because the light passes through less liquid. A cuvette is 1 cm; 200 µL in a flat 96-well plate is about 0.58 cm, so the plate reads roughly 42% lower for the same culture. Divide the plate reading by the path length in centimetres to get the cuvette-equivalent value, or use the reader's own path-length correction if it has one.
Why should I dilute a culture before reading the OD?
Because above about 0.8 the relationship between scattering and cell density stops being linear and the culture reads lower than it is. Dilute into the linear range, read, and multiply by the dilution factor — otherwise every count from that reading is an underestimate and the error grows with density.
Is OD600 the same as CFU/mL?
No. OD counts scattering particles, including dead cells and debris; CFU counts cells that divide into a colony. The two agree reasonably in log phase and diverge in stationary phase and after any treatment that kills cells without lysing them.
Related references
Related tools
Turn an OD600 reading into a cell density, a total cell count, and the volumes for a back-dilution.
How many grams to weigh out for a target molarity and volume, with the hydrate trap caught.
Convert centrifuge speed to relative centrifugal force and back, for any rotor radius.
Convert between mass, moles, molarity and copy number for DNA/RNA.
Ct values in, fold change out - with the propagated standard error, and the right method picked for your amplicon efficiencies.