Microbial / cell culture doubling time, growth rate, and projected count from two timepoint measurements.
Doubling time = elapsed × ln(2) / ln(N_t / N_0). Valid only during the exponential growth phase, before nutrients deplete or feedback inhibition kicks in.
Doubling time is the fundamental kinetic parameter of any exponentially growing population — bacteria in a flask, mammalian cells in a tissue-culture plate, yeast in a fermenter, tumor cells in vivo. Knowing it answers planning questions ("how long until I have enough cells to harvest?"), comparison questions ("is this strain growing slower than the wild type?"), and quality-control questions ("does this batch have the expected metabolism?"). The math is exact within the exponential phase: two count measurements at two timepoints uniquely determine the doubling time. Yet bench scientists routinely eyeball it from a growth curve or recall it from memory rather than computing it from their own data; this calculator removes that friction with two count fields and an elapsed-time field.
Beyond the doubling time, the calculator returns the specific growth rate μ (the natural-log-based rate constant standard in microbiology), the number of doublings observed, and a projected count at a user-chosen future timepoint — useful for scheduling subcultures or harvests. The growth curve renders the trajectory across the elapsed-plus-projection window so the exponential shape is visible at a glance.
For exponential growth: N(t) = N₀ · 2^(t / T_d) = N₀ · e^(μ·t).
Given two measurements N₀ at time 0 and N_t at time t (with t > 0 and N_t > N₀ for growth):
Projection forward from N₀ at any future time t_p is N₀ · 2^(t_p / T_d).
The math assumes (a) the population is in exponential growth phase — early stationary, lag, or death-phase data is not modeled; (b) growth conditions are constant (medium not exhausted, temperature stable, no toxin accumulation); (c) population behavior is not contact-inhibited or density-limited.
Enter initial count (cells / mL or cells / well or any consistent unit) at the start of the measurement window. Enter final count at the end of the window. Enter the elapsed time in hours. Enter a projection forward time in hours if you want a future-count estimate. The result panel shows:
The growth curve plots the predicted N(t) trajectory from t = 0 to a horizon that captures both the measurement window and the projection horizon, with a marker at the projection point.
E. coli in rich medium: N₀ = 1 × 10⁵ cells / mL, N_t = 6.4 × 10⁶ cells / mL after 6 hours. Project to t = 12 h.
In practice, by t = 12 h E. coli would be deep into stationary phase and the projection is unphysical — a useful illustration of how exponential extrapolation breaks down without the limit-of-growth context.
Yeast batch culture: 2 × 10⁵ → 1.8 × 10⁶ in 8 h.
Slow mammalian cells: 5 × 10⁴ → 2 × 10⁵ in 48 h.
Lag phase and early stationary. The formula is for pure exponential growth. If your initial timepoint is during lag (before exponential growth started), T_d is overestimated. If the final timepoint is into stationary (growth slowing), T_d is overestimated even more. Best practice: collect more than two timepoints and fit only the linear-on-log-y region.
Death phase. If N_t < N₀, the formula produces negative doubling time — biologically the population is dying. Use a death-rate model, not Riegel-style growth.
Inoculum-dependent lag. Very dilute inocula (< 10² cells / mL for bacteria) extend the lag; very dense inocula (> 10⁹ for E. coli) may already be in stationary. Check that your starting point is in mid-exponential.
Diurnal variation. Mammalian cells grown in serum can show 12–24 h cycles correlated with serum factors. A two-point fit across less than one full cycle can mislead.
Method of counting. Hemocytometer (manual) is ±10–20 % per count; automated cell counters ±5 %; flow-cytometry ±1 % but only for distinguishable populations; OD600 reads turbidity, not viable count, and saturates at ~10⁹ cells/mL. The calculator is unit-agnostic but the doubling time is only as accurate as the worse of the two counts.
Synchronized vs asynchronous cultures. Asynchronous cultures show smooth exponential curves; synchronized cultures (newly divided population) show step-like growth on fine time grids — averaging recovers exponential, but two timepoints close in time may straddle a step.
Generation time vs cell-cycle time. In microbiology these are usually synonyms. In mammalian-cell papers, "cell cycle time" can specifically mean the average time from one division to the next per individual cell, and at low growth rates it can differ from the population-doubling time because of varying fractions of dividing cells.
Death and stress effects. Stress (heat shock, antibiotic exposure) commonly extends T_d before any death is visible. Comparing T_d across conditions controls for inoculum and medium but doesn't isolate growth from death — bidirectional flow.
Substrate exhaustion. Toward the end of batch culture, the apparent T_d lengthens dramatically as carbon source runs out. A two-point fit straddling that transition gives an "effective" T_d that's not biologically meaningful.
Population vs single-cell. The formula returns the population doubling time. Individual cells in the population have a distribution of generation times around it — typically Gaussian with 10–30 % CV. Using single-cell mother-machine data requires different statistics.
Does not account for lag-time correction. Some practitioners fit a Baranyi or Gompertz model with an explicit lag and asymptote; the calculator uses the simplest pure-exponential model.