STATISTICAL CHARACTERISTICS AND ORIGIN OF OSCILLATORY ZONING IN CRYSTALS

Complex intracrystalline zoning patterns in hydrothermal garnet and ve suvianite and magmatic plagioclase were analyzed by statistical method s to test for fractal behavior. The zoning data were collected by elec tron and proton microprobe, and backscattered electron images and pola rized micrographs were digitized. The analysis shows that self-affine fractal geometry can be used to characterize the zoning patterns of ve suvianite and some garnet patterns. The range of power-law scaling ext ended up to two decades, The results from the plagioclase samples were not sufficient to determine whether or not the zoning patterns were s elf-affine. The measured Hurst exponents are mostly in the range 0.25- 0.45, indicating fractal scaling and anti-persistent behavior, This me ans that an increasing compositional trend in the past favors a decrea sing trend in the future and vice versa. No distinct periodic componen ts of the zoning patterns were found. The influence of environmental c hanges (external fluctuations) on a simple crystal growth model was in vestigated by numerical simulations. The concentration at the boundary of a diffusion layer was allowed to vary as a Brownian-motion curve, and the effect of the external fluctuation on diffusion and local grow th kinetics was investigated. We conclude that factors operating on sc ales much larger than the local interface processes are most important in controlling the zonation.