Choosey hot sand: Reflection of grain sensitivity on pattern morphology

We build and investigate a nonstandard model of pattern formation in a syst em of discrete entities evolving in discrete space and time. We chose a san dpile paradigm to fit our ideas in the frame of current research. In our mo del sand is hot because a grain can topple against gradient, i.e., the grai n can walk to another node even when a number of grains in its current node is less than a number of neighboring nodes. Sand is choosey because behavi or of the grains is not determined by any global parameter or any threshold of a number of neighboring grains (called here a grain sensitivity) but de pends on the exact number of grains in the neighboring nodes. Namely, we as sume that a grain being at a node x goes to one of the eight neighboring no des, chosen at random, if there is another grain at the node I or if the nu mber of grains in eight neighboring nodes lies in some set of 2({1,...,8}). These 256 rules of sensitivity are investigated. The classification of the rules if offered, based on the morphology of the patterns generated by eac h rule. Eight morphological classes are found. Fine structure of every clas s is investigated and transient phenomena are analyzed. Three kinds of desc ription of class rules by Boolean expressions are offered. Evolution of the classes governed by several one-dimensional parameters is considered.