Life uses physics.
Physics alone is not a definition of life.
Living systems exploit ordinary physical processes in unusually organized, sustained ways: molecules diffuse, membranes maintain gradients, chemical free energy is converted into work, patterns can self-organize, information must be copied with finite fidelity, and feedback can resist disturbances. This lab isolates those ingredients one at a time without pretending that any single mechanism explains life.
Random microscopic motion produces predictable macroscopic diffusion.
Reset every particle to the left chamber, then let thermal-like random motion spread the population. Change the diffusion step size and membrane permeability. The particles are random; the concentration trend is not.
A selective membrane can turn concentration differences into voltage scales.
For one permeant ion at equilibrium, the Nernst equation links its concentration ratio to the electrical potential that would exactly balance diffusion. Change the ion, concentrations, temperature and membrane voltage.
Directed molecular work requires an energy budget.
Use a generic stepping machine to compare chemical free energy with mechanical work against a load. The model does not represent one particular protein; it exposes the bookkeeping that every energy transducer must obey.
Local reactions plus diffusion can create structure without a central designer.
The Gray-Scott reaction-diffusion equations are a classic nonlinear toy system. Seed a local disturbance and change feed and removal parameters to watch homogeneous conditions destabilize into persistent spatial patterns.
Long sequences turn tiny per-symbol error rates into a systems problem.
Choose a sequence length and copying error probability. The chance of an exact copy falls as errors accumulate. Then turn on a simple three-copy majority code to see how redundancy can suppress independent errors.
Negative feedback can keep a variable near a target while the environment pushes it away.
A simple regulated reservoir has baseline input, first-order loss and a controller that increases or decreases net actuation according to the error from a setpoint. Apply a sustained disturbance and change feedback gain.