The universal solution for this kind of problem is to "normalize" all the data. Every dataset, then, would be rescaled to a span of 0.0 to 1.0 - typically by subracting from each value the min value that you want to represent in your visualization (frequently 0) and dividing each result by the real span (max(data) - thatMinValue). After such a treatment, two datasets, normalized separately, one with a max value of 20, the other with a max value of 40, could be visualized in the same dataspace with range of 0.0 to 1.0 in all directions, and the 20 magnitude vector in the first dataset would display with the same magnitude as the 40 magnitude vector in the second dataset. (IDL Object Graphics has a routine set of keywords [XYZ]COORDCONV for all its plottable graphics objects, which is specifically designed to aid in this kind of normalization. In IDL Direct Graphics, however, you have to create more of your own variables to manage the scaling and mapping of the different graphics elements.) James Jones