Classification of Semitotalistic Cellular Automata in Three Dimensions
Carter Bays
University of South Carolina, Columbia, SC 29208, USA
Abstract
This paper describes a mechanism by which three-dimensional semitotalistic cellular automata (CA) may be classified. The classification scheme is based upon the beahavior of specific CA rules when originally configured (a) as isolated forms and (b) as random "primordial soup." The classifications of the behavior under these two separate schemes may then be orthogonally related.
Most of the simulations described herein were done in a universe of dense-packed spheres, where each cell has 12 touching neighbors. Results also apply to 
, with 26 neighbors touching each (cubic) cell.
