The need for accurate wood quality prediction in hardwoods has increased with the growing importance of these species to the global forest industry, in particular Eucalyptus species. Several process-based models of xylem development have been produced (Wilson 1964; Deleuze and Houlier 1998; Fritts et al. 2005) but have considered woody tissue development in softwoods only. The situation is more complex in hardwoods, which produce a greater variety of cell types in the xylem. The model described here, which we have called "CAMBIUM", simulates the growth and development of a hypothetical, simulated population of cells in variable numbered cell files, bounded within sets of rays. Although models for softwood xylem effectively work in a single dimension (modeling a single radial file of cells), it is necessary to consider hardwood xylem in at least two dimensions in the transverse plane (modeling multiple, neighboring files of cells) because of the highly significant interactions between cell types and the importance of their spatial arrangement. These interactions are constantly invoked in the hypothetical cell population during the model run to modify environmentally determined "potential" cell growth and development. The model predicts cell fate, cell division and production, cell growth, and cell secondary wall development in individual cells in the simulated population. It uses outputs of whole-tree physiological parameters generated on a daily time-step from the CABALA process-based model (Battaglia et al. 2004) which has been widely tested and validated for various Eucalyptus spp. Outputs of particular importance from CABALA are daily derived estimates of pre-dawn and midday water potential, stem and foliage allocation, minimum and maximum temperature, and vapour pressure deficit. CAMBIUM relies heavily on conceptual morphogenic fields of growth-regulating substances, particularly auxin, which have been found to affect cambial activity and cyto-differentiation in the cambial zone and developing xylem. The model considers different approaches to determining cell fate as either a function of cell position with these morphogenic fields or as a function of the flow of chemical signals, or as a function of both. It currently determines between fibers or vessels in the developing xylem from a homogenous population of cambial initials. The model also provides more than one approach to calculating the potential daily growth of developing cells. Following the additional consideration of cell-cell interactions, the model is thus capable of generating high-resolution estimates of fiber and vessel transverse dimensions. The process of secondary wall thickening is currently very simple, similar to that proposed by Deleuze and Houllier (1998) and generates estimates of fiber and vessel wall thickness. Wood density is derived empirically from cell transverse and wall dimensions as well as potentially from vessel frequencies. Model parameterization and validation is currently underway using data from four Eucalyptus species from three separate research trials in South Africa and Australia. Early results indicate that the model is able to accurately detect changes in wood properties in a radial profile in response to changing environmental conditions in a forest stand.