A leaf can look dry while still holding droplets large enough to shelter bacteria. The missing water has not necessarily vanished everywhere: it may remain in isolated patches too small for us to see. Whether a cell survives depends partly on its immediate surroundings, not simply on whether the whole leaf looks wet.
That distinction separates avoiding complete drying from tolerating the loss of water. It also explains why a count of bacteria across an entire leaf can miss something important about their prospects.
Tiny droplets need the right conditions
In experiments on microscopic surface wetness, researchers dried bacterial suspensions on artificial surfaces under controlled humidity. As the liquid retreated, its edge caught around cells and aggregates, leaving tiny droplets. Dissolved substances that attract moisture helped maintain the remaining liquid through deliquescence: absorbing water until they dissolve in it.
Larger bacterial aggregates tended to have larger droplets, and survival was higher in larger droplets. This gives a physical explanation for a refuge on a surface that appears dry.
The experimental surface was not a living leaf. The result supplies a mechanism relevant to leaf habitats, rather than proof that every apparently dry leaf contains identical droplets. Humidity and dissolved substances matter. Bacteria cannot conjure liquid from arbitrarily dry air, and a concentrated solution is not equivalent to fresh rainwater.
Clusters change who survives
A separate bean-leaf study followed Pseudomonas syringae strain B728a under moist conditions and repeated drying. Under moist conditions, living cells were common whether they were solitary or aggregated. Drying made the difference between those arrangements much more consequential.
Seven days after inoculation, following repeated drying, more than 90% of the living cells were in aggregates containing at least 100 cells. Read the denominator carefully: this does not mean that more than 90% of all bacteria survived. It describes where the survivors were found.
The study supports preferential survival in clusters. It does not require a story in which bacteria consciously choose a safe settlement. A clustered population can emerge as cells multiply beside one another and exposed cells survive less well. Its final pattern records both growth and losses.
These two experiments answer complementary questions. The leaf experiment tracks survival in a plant habitat; the artificial-surface experiment makes tiny water refuges easier to examine. Neither alone establishes every step of the other.
Green does not mean food everywhere
Water is only part of the problem. Leveau and Lindow’s sugar-reporter experiment used engineered bacteria on bean leaves to investigate access to sugars. A genetic reporter produced green fluorescent protein in response to fructose consumption, with sucrose also eliciting a response.
The researchers found strong variation among cells and over time. A leaf’s ability to make sugars does not give every surface resident continuous access to them.
The fluorescence was an experimental signal, not the natural color of those microbes. Nor was one cell’s brightness a direct sugar-concentration meter: the authors explain why the signal needs interpretation in relation to other cells, time, and consumption. That distinction matters when a vivid microscope image seems to offer an immediate map of food.
A head count needs a map
Imagine two leaf samples with the same number of living bacteria. One contains mostly exposed individuals; the other contains clusters. The bean-leaf results show why those arrangements can have different consequences during drying. Equal totals need not mean equivalent habitats.
Our article on aquatic animals in moss follows water films at another scale, including animals that can become dormant. Here, the central mechanism is more local: a surviving bacterium may still occupy a wet refuge after visible water disappears. “Dry” describes what the observer can detect; explaining survival requires asking what conditions reach the cell.
