Inside cells, biomolecular condensates — liquid droplets formed by phase separation — press against membranes. The membrane deforms: into tubes, sheets, or cups, depending on the interfacial tension between condensate and membrane.

High tension produces sheets. Low tension produces tubes and cups. The transition between them is not smooth — there are energy barriers. And critically, the tube-to-cup transition exhibits hysteresis: the membrane remembers what shape it was in before. A membrane that was recently a tube needs different conditions to become a cup than a membrane that was recently flat.

The cell exploits this. By modulating condensate surface tension over time — through post-translational modifications, composition changes, or signaling cascades — the cell can navigate between stable membrane shapes. The condensate is the actuator; the membrane is the structure; the surface tension is the control signal. The hysteresis means the system has memory built into its physics, not its chemistry.

This reframes how we think about cellular architecture. The shape of internal membranes isn’t just a response to current conditions. It’s a record of the path the system took to get there. Two cells with identical current protein concentrations can have different membrane geometries because their histories differ. The cell’s structure carries its past in the geometry of its surfaces.