Squirm3 > EvSelfReps > membrane

Evolvable Self-Replicating Molecules in an
Artificial Chemistry

Replication within a membrane

Below the experiment is running as a java applet - if you don't see anything then you either need to change your browser settings to enable java, or use a different browser.

You can control how fast the applet runs by using the slowdown controls, if enabled they cause a delay of so many milliseconds per timestep. Remember to hit 'update applet' after changing the value.

If you want maximum speed then there are controls that let you draw only every 100 timesteps, for example. If you really want maximum speed then take a look at the C++ version on the development page.

The 'pause/run' button is a toggle switch that lets you stop and start the applet.

When you hover over the atoms with the mouse their type and state are displayed. This works best when the applet is paused.

To reset the applet, use your browser's Refresh button.

Slowdown: milliseconds
Draw only every: time steps


When the applet starts there is an e8-a1-f1 molecule in a random soup. It is locked inside a loop of a4's - these do not readily react with other atoms and are effectively inert for our purposes here. The molecule can move freely but cannot pass through the membrane because the physics does not allow bonds to stretch greater than a single diagonal. However, single atoms can pass through the membrane freely across diagonal links and so the molecule inside is able to replicate.

After it has replicated several times, however, it runs out of room since the membrane does not grow. Eventually all replication grinds to a halt.

Some mechanism that enabled the membrane to grow would help the replicators inside. Such an effect could be viewed as being beneficial to the membrane itself, or you could view the membrane and the replicator as a single entity...

The java source for this applet is here: sq3.zip (13k) - for instructions on building the source see the development page.