The internal zonation of minerals is commonly observed in metamorphic and igneous rocks and has traditionally been ascribed to diffusion or growth processes. These zones have been used to study the rates of diffusion, rates of cooling, and changes in melt and grain boundary chemistry.Most studies of zoning concentrated on porphyroblasts [Spear et al., 1984; Hickmottet al., 1987; Halden & Hawthorne, 1993], however recent studies suggest that even in matrix grains we can learn a lot from these zoning patterns. It has been pointed out that grain boundary migration processes can also lead to an observable mineral zonation [Reinecke et al. 2000; Holness & Watt 2001;Jessell et al 2003]. Clearly it is of some importance that we can distinguish between these processes if we are to correctly interpret them.

Modern electron microprobe mapping and profiling techniques can routinely reach spatial resolutions of 1-3mm for major elements, and there is no sign at the present time that trace element mapping has reached it resolution limit. We thus have the prospect in the coming years of being able to map out zoning patterns in matrix grains for a wide variety of elements, which will greatly improve our ability to both constrain the fundamental parameters of diffusion, grain boundary mobility etc, and simultaneously allow us to unravel quite complex geological histories.

In this paper we describe a series of simple numerical experiments that each produce a zonation resulting from a single process, and compare the resulting patterns as the basis for a set of criterion that help us to distinguish the processes.