>I use an algorithm that manipulate monads.
>
>I begin with 50+ randomly placed monads.
>
>They always seem to break into groups of two.
>
>The rules, the initial conditions, and some of the results I've gotten are accessible here:
>
>http://cloudmusiccompany.com/science/research.htm
>
>My question is, why do the monads split into clusters of two?

My gut reaction is that it's a byproduct of VFP's implementation of RAND. It would be interesting to see this redone with a cryptographic random number (which should normally be evenly distributed on a graph if it's truly random) vs. VFP's version.