The Bacterial Growth Curve Is One Equation, Not Four Rules

Four phases, one underlying cause each

Plot the log of a bacterial population against time in a closed batch culture — a flask with a fixed volume of nutrient broth, no feeding, no removal — and you get the same characteristic S-shaped curve every microbiology student learns: lag, then exponential (log) growth, then stationary, then death. What is less often taught is that all four phases fall directly out of one underlying mechanism — nutrient depletion and waste accumulation acting on a single growth-rate equation — rather than being four separate biological programs the cells switch between.

The Monod equation: growth rate as a function of what's left to eat

Jacques Monod's 1949 equation is the mathematical engine behind the whole curve. It says the specific growth rate μ (the fractional rate of population increase per unit time) is not a fixed constant — it depends on the current nutrient concentration S, saturating toward a maximum rate as nutrient becomes abundant and collapsing toward zero as it runs out:

mu(S)  =  mu_max * S / (Ks + S)

mu_max = maximum possible specific growth rate (nutrient never limiting)
Ks     = half-saturation constant — the nutrient concentration at which
         growth rate is exactly half of mu_max (a measure of how "hungry"
         the organism is for scarce nutrient; smaller Ks = better scavenger)

dN/dt  =  mu(S) * N              // population grows proportional to itself,
                                  // scaled by the current growth rate
dS/dt  =  -(1/Y) * mu(S) * N     // nutrient is consumed proportionally to
                                  // growth, divided by yield coefficient Y

Y, the yield coefficient, is how many new cells a given amount of consumed nutrient produces — a fixed conversion efficiency. This is genuinely a coupled system: population N depends on the growth rate, the growth rate depends on nutrient S, and nutrient depends on how much the population has already eaten. That coupling, run forward in time, is what generates every visible feature of the classic growth curve without needing any additional rule.

Lag phase: not dormancy, but metabolic retooling

When cells are transferred into fresh medium, the population size barely moves for a while even though nutrient is abundant and μ(S) from the Monod equation alone would predict immediate fast growth. The missing piece is that Monod's equation assumes the cells are already metabolically ready to exploit the nutrient in front of them — freshly transferred cells typically are not. If the new medium contains a different carbon source, or the previous culture was nutrient-starved or in stationary phase, cells need time to synthesize the specific enzymes and transporters required (a process controlled largely by the stringent response and associated regulatory systems that were suppressing growth-related gene expression under the previous, harsher conditions), repair accumulated macromolecular damage, and rebuild ribosome content before division machinery can run at full speed. This is why lag phase length is highly history-dependent: cells transferred from a similar, nutrient-rich medium show almost no lag at all, while cells transferred from starvation or a very different substrate can lag for hours — the model captures this with an explicit delay or a slowly-rising effective mu_max term, since the bare Monod equation on its own has no natural lag built in.

Exponential phase: μ pinned at its maximum

Once retooled and while nutrient S remains far above the half-saturation constant Ks, the Monod term S/(Ks+S) sits close to 1, so μ ≈ μ_max — a constant — and the population equation dN/dt = μN integrates to simple exponential growth, N(t) = N0 * e^(μ_max * t). This is the textbook phase where population doubles at a fixed interval, the doubling time td = ln(2)/μ_max, famously around 20 minutes for E. coli under ideal lab conditions (37°C, rich medium) — one of the fastest generation times of any cellular organism, though most environmental bacteria grow many times slower under real-world nutrient limitation and temperature stress.

Stationary phase: the Monod term itself collapsing

As S is consumed, once it drops down toward and below Ks, the fraction S/(Ks+S) is no longer close to 1 — growth rate genuinely slows, continuously, not as a sudden switch. Population growth decelerates and eventually flattens as net growth rate approaches zero: new cell division roughly balances a background rate of cell death, so the population size N stays approximately constant even though individual cells are still dividing and dying underneath the flat curve. Two mechanisms compound the slowdown beyond simple substrate depletion: accumulating waste products (acidic fermentation byproducts, autoinducers, and other metabolic waste) actively inhibit growth directly, and rising cell density triggers quorum sensing — bacteria detecting their own population density through secreted signal molecules and switching gene expression toward stress-survival programs (thicker cell walls, dormancy-like states, sporulation in species capable of it) rather than continued rapid division, even before nutrient is fully exhausted.

Death phase: when maintenance can no longer be paid for

Eventually nutrient is too scarce and waste too concentrated for cells to maintain basic membrane integrity and repair processes, and the population declines — often, empirically, following its own roughly exponential decay curve, N(t) = N(t_stationary) * e^(-k_d * (t - t_stationary)), with a death-rate constant k_d that is typically much smaller in magnitude than the growth phase's μ_max, which is exactly why the death phase visually stretches out over a much longer, gentler decline than the sharp climb of exponential phase. A subtlety worth knowing: death phase is rarely perfectly exponential in practice, because a genetically and physiologically heterogeneous population does not die uniformly — some subpopulation of persister cells with lower metabolic activity survives far longer than the bulk average, which is one of the mechanisms behind antibiotic tolerance (as distinct from resistance) in clinical and industrial settings.

Why this one model, run once, produces the whole shape

The elegance of the Monod-kinetics simulation on this page is that lag, log, stationary and death are not four separately coded rules — they are one continuously running system (population, nutrient and, in a fuller model, waste concentration, all coupled through their differential equations) that simply passes through four visually distinguishable regimes as the ratio of S to Ks moves from far-above to far-below 1, with waste accumulation and cell-death terms shaping the tail. That is the same principle behind nearly every population model on this site: a small number of coupled differential equations, integrated forward in time, reproducing complex-looking qualitative behaviour without complex-looking rules.

Frequently asked questions

Try it live

Everything above runs in your browser — open Bacterial Growth Curve and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.