An array in which an external control feature for normalization has been designed and tested for its ability to mimic the range of observed expression levels for a test set of oligos. The external control probes span a series of concentrations. They are spatially randomized across a grid of an array. The series of concentrations is duplicated in a given grid. The individual grid layout and number of control and external normalization features per grid have been designed to cope with sources of both systematic error and spatial variation.