Drop E. coli into a flask of fresh food and its numbers rise and fall along a predictable S-shaped path called the growth curve. The curve is usually split into four phases: a pause, rapid doubling, a plateau, and a decline.
The Four Phases at a Glance
- Lag phase: cells adjust to the new food and barely increase in number. In one lab culture this took about 1.75 hours.1Chinese Journal of Chemical Engineering. Kinetics of High Cell Density Fed-batch Culture of Recombinant Escherichia coli Producing Human-like Collagen
- Log (exponential) phase: cells divide at a steady rate, so the population doubles again and again.
- Stationary phase: once the sugar is used up, the cells’ energy reserves start to fall and growth stops.2PubMed. On characterization of the growth of Escherichia coli in batch culture
- Death phase: the number of living cells drops, typically at a steady proportional rate rather than all at once.3PubMed. Death Rate of E. coli during Starvation Is Set by Maintenance Cost and Biomass Recycling
Cell numbers multiply rather than add, so graphs plot them on a logarithmic scale. That turns steady doubling into a straight line, and the steepness of the line tells you how fast the cells are dividing.
On a textbook graph, the x-axis is time and the y-axis is usually the log of cell number or cloudiness. Many class experiments stop before the death phase shows up.
These phases describe a closed flask. The flask gets food once and nothing is removed, so waste builds up and the food runs out. In the gut, or in industrial tanks with steady feeding, cells can keep growing far longer.
Log phase is fast. Under good lab conditions E. coli can double roughly every 20 minutes, which means one cell could become over a million in about seven hours.
In a real flask, the boundaries between phases are gradual. Where a textbook draws a corner, a real curve rounds off, so two people reading the same graph may place a boundary at slightly different times. A flat line alone doesn’t tell you the phase: a flat stretch at the start is lag, while a flat stretch after a rapid rise is stationary.
What Happens During the Lag Phase
Cells start adjusting within about 4 minutes of being placed in fresh broth, first switching on genes for taking up phosphate.4PubMed Central. Lag Phase Is a Distinct Growth Phase That Prepares Bacteria for Exponential Growth and Involves Transient Metal Accumulation Around 20 minutes in, a main program begins that turns up 945 genes for building proteins, making DNA building blocks, and using oxygen for energy.5PubMed Central. Lag Phase Is a Distinct Growth Phase That Prepares Bacteria for Exponential Growth and Involves Transient Metal Accumulation Cells also briefly soak up metals such as iron during this time.6PubMed Central. Lag Phase Is a Distinct Growth Phase That Prepares Bacteria for Exponential Growth and Involves Transient Metal Accumulation7PubMed Central. Lag Phase Is a Distinct Growth Phase That Prepares Bacteria for Exponential Growth and Involves Transient Metal Accumulation
Inside the lag there are two steps. At first, cells mainly switch on the enzymes needed to use the specific sugar in the medium, plus a gene for handling oxidative stress.8PubMed Central. Promoter activity dynamics in the lag phase of Escherichia coli They still don’t measurably grow in size. Once cells start to enlarge, they flip to a full growth program, including the protein-making machinery.9PubMed Central. Promoter activity dynamics in the lag phase of Escherichia coli
How long the lag lasts depends on where the cells came from. Cells taken from an old, starved culture need more repair time than cells from a growing one. Moving cells into a very different food means building many new enzymes, while moving them into the same food they were already using can make the lag almost vanish. A big change in temperature adds adjustment time too.
Not every cell lags the same amount. Watching single cells shows wide variation in how long each one waits. In one strain over 90% of cells showed no lag at all, while in a strain with many genes removed, about 80% waited 0.5 to 3 hours. The lag of a whole culture is an average of these individual waits.
Exponential Growth in the Log Phase
Only part of the log phase is truly balanced, when cell number, cell volume, cloudiness and energy content all rise at the same rate.10PubMed. On characterization of the growth of Escherichia coli in batch culture In one culture that perfectly even window lasted just about 1.5 hours.11PubMed. On characterization of the growth of Escherichia coli in batch culture Lab protocols that call for “mid-log” cells are aiming for that window. Scientists harvest cells in log phase because the cells are similar to each other, which gives repeatable results.
Speed depends on the food. Growth rate rises with the amount of available food, but levels off once food is plentiful.12PubMed Central. A brief history of bacterial growth physiology When growing more slowly, cells keep each protein-making unit working at full speed but use fewer of them.13PubMed Central. Reduction of translating ribosomes enables Escherichia coli to maintain elongation rates during slow growth
You can calculate doubling time from the graph. Pick two points on the straight part of the line and see how long the count takes to double. A shorter doubling time means faster growth. Points from the curved parts give misleading answers.
Why Growth Stops in Stationary Phase
When the sugar is exhausted, the cell’s energy reserves begin to fall, marking the start of stationary phase.14PubMed. On characterization of the growth of Escherichia coli in batch culture A master stress switch rises quickly and turns on a large set of survival genes.15PubMed Central. Stress sigma factor RpoS degradation and translation are sensitive to the state of central metabolism It drops back to low levels when growth resumes.16PubMed Central. Stress sigma factor RpoS degradation and translation are sensitive to the state of central metabolism
In a typical lab flask the limiting nutrient is the carbon source, such as the sugar, though oxygen can also become limiting in a dense, shaken culture. Because the stress switch rises as the sugar disappears, the cell reprograms itself before it is completely starved, which gives it a head start. The count stays flat even though some cells may still divide while others die, and waste products and acid build up alongside the loss of food.
The cells change shape too. Normal E. coli become small and round in stationary phase instead of rod-shaped.17PubMed Central. Growth phase-regulated expression of bolA and morphology of stationary-phase Escherichia coli cells are controlled by the novel sigma factor sigma S Cells lacking this stress switch stay rod-shaped and larger.18PubMed Central. Growth phase-regulated expression of bolA and morphology of stationary-phase Escherichia coli cells are controlled by the novel sigma factor sigma S Starving cells shrink further by breaking down their own reserves and parts of their inner wall.19PubMed Central. Survival guide: Escherichia coli in the stationary phase They then build a toughened outer wall that withstands harsh conditions.20PubMed Central. Survival guide: Escherichia coli in the stationary phase That thicker wall resists heat, acid and chemicals better, which is why old cultures can survive treatments that kill log cells.
Aged cultures can even evolve. Cells kept in old cultures can gain mutations that let them outcompete fresh cells.
The Death Phase and Cells That Only Seem Dead
During starvation, survival drops at a steady proportional rate.21PubMed. Death Rate of E. coli during Starvation Is Set by Maintenance Cost and Biomass Recycling Surviving cells feed on nutrients released from dead cells.22PubMed. Death Rate of E. coli during Starvation Is Set by Maintenance Cost and Biomass Recycling Some dead cells are actively broken open by a cell program, releasing proteins that can feed the next generation.23PubMed. Function of the sigma(E) regulon in dead-cell lysis in stationary-phase Escherichia coli
That recycling is why the death phase rarely ends with every cell dead. A small population can persist for a very long time. On a graph, the decline is often steep at first and then flattens into a long, slow tail rather than dropping straight to zero. The survivors aren’t necessarily the same as the cells that began stationary phase, since some have changed in ways that help them live in the old, depleted culture.
Many cells that look dead aren’t. Early in stationary phase, many E. coli become viable but nonculturable (VBNC): alive, but unable to form colonies on a standard plate.24IntechOpen. Survival Strategy of Escherichia coli in Stationary Phase: Involvement of σE-Dependent Programmed Cell Death Plate counts can therefore underestimate how many cells are really alive.25PubMed. Detection of Viable but Nonculturable E. coli Induced by Low-Level Antimicrobials Using AI-Enabled Hyperspectral Microscopy Mild stresses such as low-level disinfectants, acid, salt or starvation can push E. coli into this state.26PubMed. Detection of Viable but Nonculturable E. coli Induced by Low-Level Antimicrobials Using AI-Enabled Hyperspectral Microscopy Some cells can later recover and grow again, keeping traits like antibiotic resistance.27PubMed. Viable but non-culturable state formation and resuscitation of different antibiotic-resistant Escherichia coli induced by UV/chlorine
For food and water safety, this means a test showing zero colonies doesn’t guarantee zero living bacteria. Disinfectants at too-low doses may push cells into hiding rather than kill them, and recovered cells can still cause harm.
How the Growth Curve Is Measured
The quickest method shines light through the culture and records how cloudy it is, a reading called optical density, or OD. This reading counts all bacteria, dead and alive.28PubMed. Escherichia coli viability determination using dynamic light scattering: a comparison with standard methods At higher readings, cloudiness stops rising in step with cell numbers.29PubMed Central. Estimating microbial population data from optical density For dense cultures, dilute the sample before reading.
The reader detects light scattered by anything in the broth, so clumps of cells, debris, and dead cells all raise the reading as much as healthy ones do. Cell size matters too: small stationary cells scatter less light than larger log cells, so the same reading can stand for different numbers of cells. Cloudiness readings are best used to compare within one experiment rather than to count cells exactly.
Plate counts work differently. You spread a diluted sample on a plate and count colonies the next day. Only cells that can grow are counted, so the method is slow but separates living from dead. It misses the alive-but-won’t-grow-on-a-plate cells described above.
For a class growth curve lab, take readings at even intervals, more often during log phase. Use a blank of plain broth to zero the reader, and plot the log of the readings so log phase shows as a line. Note the temperature and broth type with your data.
What Changes the Shape of the Curve
E. coli’s maximum growth speed changes with temperature, with lab growth measured from 27 to 45°C.30PubMed Central. Data on the influence of temperature on the growth of Escherichia coli in a minimal medium containing glucose as the sole carbon source for the joint computation of growth yields and rates at each temperature from 27 to 45°C31PubMed Central. Temperature-dependent growth kinetics of Escherichia coli ML 30 in glucose-limited continuous culture How well cells grab scarce sugar stays about the same across temperatures.32PubMed Central. Temperature-dependent growth kinetics of Escherichia coli ML 30 in glucose-limited continuous culture At high temperatures growth and division become irregular, but cells pre-adapted to heat grow smoothly.33PubMed. Dynamics of Escherichia coli at elevated temperatures: effect of temperature history and medium A growth curve is only comparable between experiments when the temperature was the same. The usual lab setting is 37°C, human body temperature.
Food changes the curve as well. Given glucose and lactose together, E. coli uses glucose first, pauses, then switches to lactose, producing a two-step curve. The pause is a second lag while cells build lactose-using enzymes. Glucose blocks lactose use partly by keeping the lactose genes switched off through a repressor, not just through a signal molecule as textbooks often say.34PubMed. Mechanism responsible for glucose-lactose diauxie in Escherichia coli: challenge to the cAMP model
To read a two-step curve, look for two straight rising stretches separated by a flat shoulder. The first slope is usually steeper because glucose supports faster growth, and the shoulder’s length shows how long switching takes.
The broth matters too. Rich broth like LB gives fast early growth, while defined simple media give more predictable curves. In one production setup, a modified simple medium gave about twice the cell mass of LB.