The Bacterial Growth Curve

Why the growth curve matters

• It is the founding measurement of microbial physiology. Almost everything we know about how fast microbes grow, what limits them, and how they adapt comes from watching a flask move through these phases. Jacques Monod's 1949 Annual Review of Microbiology paper "The growth of bacterial cultures" turned the curve into equations relating growth rate to substrate concentration, and won part of the intellectual scaffolding that led to his 1965 Nobel Prize for the operon.
• Antibiotic susceptibility depends on phase. Beta-lactams like penicillin kill only actively dividing cells because they block cross-linking of new peptidoglycan at the septum — so log-phase cultures are exquisitely sensitive while stationary-phase cells shrug them off. This phase-dependence is a leading reason chronic and biofilm infections resist treatment.
• Industrial fermentation is timed to the curve. Primary metabolites (ethanol, lactic acid, most amino acids) accumulate during exponential growth, whereas many secondary metabolites — penicillin, streptomycin, and other antibiotics — are made in stationary phase (the idiophase). Knowing which phase a product tracks tells a bioprocess engineer when to harvest.
• Food safety runs on lag and log. Predictive microbiology models a pathogen's lag time and doubling time to estimate how long food is safe at a given temperature. Refrigeration works largely by lengthening lag phase and stretching the generation time of organisms like Listeria and Salmonella.
• Diagnostic turnaround is a doubling-time problem. Because Mycobacterium tuberculosis doubles only every 15 to 20 hours, a visible colony takes weeks, which is why TB culture is slow and why rapid molecular tests were such a leap. Fast growers like E. coli give an answer overnight.
• Persisters hide in the plateau. A tiny, non-growing subpopulation that arises especially in stationary phase survives antibiotic courses and reseeds infection. Understanding the stationary-phase state is central to tackling relapsing and recurrent disease.

How the growth curve works, phase by phase

Lag phase. When a small inoculum is transferred into fresh medium the viable count stays flat — but the cells are working hard. They repair damage carried over from stationary phase, rebuild ATP and ribosome pools, and induce the specific transporters and catabolic enzymes the new substrate demands. A cell moving from glucose to lactose, for instance, must synthesize the products of the lac operon before it can grow. Lag length scales with how old and how stressed the inoculum is and how different the new medium is; mid-log cells transferred into identical medium show almost no lag.

Exponential (log) phase. Now growth is balanced: every cellular component is synthesized at the same relative rate, and the population divides by binary fission at a constant maximal rate. Numbers double each generation — 1, 2, 4, 8, 16 … — so cell count rises geometrically while its logarithm rises linearly. The instantaneous rate obeys dN/dt = µN, whose solution is N = N₀·e^{µt}; the specific growth rate µ and the doubling time g are linked by g = ln 2/µ ≈ 0.693/µ. For E. coli in rich broth at 37 °C, g can be about 20 minutes. Fast-growing bacteria cheat the constraint that chromosome replication takes ~40 minutes by firing multiple replication forks per origin (multifork replication), so a cell can divide faster than it can copy its DNA start to finish.

Stationary phase. No closed flask sustains exponential growth for long. As the limiting nutrient — usually the carbon and energy source, but sometimes nitrogen, phosphate, iron, or oxygen — is depleted, and as fermentation acids, ethanol, and other end products accumulate and shift the pH, the division rate falls until it exactly equals the death rate. Net growth is zero and the curve flattens, typically near 10⁹ to 10¹⁰ cells/mL. Cells are not idle: in E. coli the alternative sigma factor RpoS (σˢ) reprograms roughly a tenth of the genome into a general stress response — thickening the envelope, inducing catalase, ppGpp-driven stringent response, DNA-protective Dps, and the protectant trehalose — making stationary cells far tougher than log cells. This is also the density window in which quorum-sensing autoinducers cross threshold, coordinating biofilm formation, sporulation, competence, and virulence.

Death (decline) phase. With reserves gone and the environment hostile, viable cells die exponentially — the log of the viable count now falls in a straight line — as autolysins, oxidative damage, and pH extremes take their toll. Death is rarely total: a resistant fraction can persist for months, feeding on nutrients released by lysing neighbors (cryptic growth). Many cultures reach a slowly evolving long-term stationary or "prolonged decline" phase in which successive genetically fitter mutants take over — the GASP phenotype (growth advantage in stationary phase), first reported in E. coli by Zambrano and colleagues in Roberto Kolter's lab in 1993 and later dissected in detail by Steven Finkel.

Doubling time and generation time — the arithmetic

Generation time (doubling time), g, is the time for the population to double during balanced growth. Take two viable counts, N₀ and N, separated by time t. The number of doublings is n = (log₁₀ N − log₁₀ N₀)/log₁₀ 2 = 3.3·(log₁₀ N − log₁₀ N₀), and the generation time is simply g = t/n. Equivalently, using the specific growth rate µ from N = N₀·e^{µt}, g = ln 2/µ. A culture that grows from 10³ to 10⁹ CFU/mL in 200 minutes has undergone 3.3 × 6 ≈ 20 doublings, so g ≈ 10 minutes. The table below shows how wildly g ranges across the microbial world.

Organism	Typical doubling time	Condition / note
Vibrio natriegens	~10 min (< 10 min reported)	Fastest-known free-living bacterium; a lab workhorse
Escherichia coli	~20 min	Rich medium (LB), 37 °C; hours on minimal salts
Bacillus subtilis	~25–30 min	Rich medium; sporulates when starved
Pseudomonas aeruginosa	~30–60 min	Depends heavily on medium
Mycobacterium tuberculosis	~15–20 h	Why TB culture takes 2–6 weeks
Mycobacterium leprae	~12–14 days	Cannot be grown in axenic culture at all
Deep-subsurface bacteria	Estimated centuries–millennia	Energy-starved sediments and rock

Batch culture vs continuous culture (chemostat)

The classic growth curve is a portrait of a batch culture: a closed flask in which nothing is added and nothing is removed, so the environment inexorably changes and the population must march through all four phases. A continuous culture escapes that fate by turning the flask into an open, steady-state system.

Property	Batch culture	Continuous culture (chemostat)
System type	Closed — fixed medium	Open — medium in, culture out at equal rate
Nutrients	Finite, depleted over time	Continuously replenished; one held limiting
Waste	Accumulates in the flask	Washed out continuously
Growth phases	Full lag → log → stationary → death	Held at a fixed, steady exponential-like state
Growth rate	Varies through the run	Set by the operator via dilution rate
Cell density	Rises to a plateau then falls	Constant at steady state
Introduced	Implicit since 19th-century culture	Monod and Novick–Szilard, 1950
Main use	Diagnostics, teaching, many fermentations	Physiology at defined rate, bioprocessing

Common misconceptions

• "Lag phase means the cells are dormant." The count is flat but the cells are metabolically frantic — repairing damage and inducing the enzymes and ribosomes the new medium requires. Dormancy (true metabolic shutdown, as in spores or persisters) is a different state entirely.
• "Exponential growth can go on forever if you feed the culture." Even a continuous culture only mimics exponential phase; in any real closed flask, space, oxygen transfer, and waste toxicity impose limits long before the medium could be infinitely refreshed. Unbounded exponential growth is a mathematical idealization.
• "Stationary phase means the cells stopped doing anything." Net growth is zero because birth and death balance, not because metabolism stops. Stationary cells are actively running the RpoS stress program, making them tougher and, in many species, initiating sporulation, competence, or antibiotic synthesis.
• "Optical density is a cell count." OD₆₀₀ reports turbidity — total biomass including dead cells and debris — and is non-linear above roughly OD 0.4. That is exactly why OD stays high in death phase while the viable plate count (CFU/mL) collapses: dead cells still scatter light but form no colonies.
• "The curve is the same for every organism and medium." Phase durations, the height of the plateau, and the doubling time all depend on the strain, the medium, temperature, aeration, and inoculum age. The shape is general; the numbers are not.
• "Binary fission just splits a cell in half at random." It is a precisely orchestrated event: the tubulin homolog FtsZ polymerizes into the Z-ring at midcell, recruits the divisome, and constricts in coordination with chromosome replication and segregation so each daughter gets one complete genome.

Famous experiments and history

• Monod's growth kinetics (1942–1949). Jacques Monod's doctoral work and his 1949 review "The growth of bacterial cultures" defined exponential and stationary phases quantitatively and introduced the Monod equation µ = µₘₐₓ·S/(Kₛ + S), which relates growth rate to limiting-substrate concentration — the microbial analog of Michaelis–Menten kinetics.
• Diauxie (Monod, 1941). Growing bacteria on a mix of two sugars, Monod saw a two-step curve with a growth pause in the middle: cells consume the preferred sugar (glucose), pause to induce the enzymes for the second (e.g. lactose), then resume. This diauxic pause helped inspire the operon model of gene regulation Monod later built with François Jacob (1965 Nobel Prize).
• The chemostat (Monod; Novick and Szilard, 1950). Independently, Monod and the team of Aaron Novick and Leo Szilard described continuous culture, freeing physiologists from the moving target of the batch flask and enabling mutation-rate and selection experiments at a fixed growth rate.
• Quorum sensing in Vibrio fischeri (Nealson, Platt, Hastings, 1970). Bioluminescence in this squid symbiont switches on only at high cell density; the discovery of the diffusible autoinducer (an acyl-homoserine lactone) and the LuxI/LuxR system revealed that bacteria count themselves — the density signal peaks precisely as cultures enter stationary phase.
• GASP and long-term stationary phase (Zambrano et al., 1993; Finkel, 2000s). Aging E. coli cultures do not simply die off: fitter mutants repeatedly sweep the population in the growth-advantage-in-stationary-phase (GASP) phenomenon — first reported from the Kolter lab in a 1993 Science paper and later reviewed and extended by Steven Finkel — showing that even a "dying" flask is a dynamic evolutionary arena.