Current schemes for hydrometeor classification are based on linear polarization measurements such as Z_h, Z_dr, LDR, ρ_co and K_dp. However, it is well known that the linear polarimetric measurables depend on particle shape and orientation in a complicated manner which tends to introduce ambiguity in classification of particle types. Since the full 3X3 polarimetric covariance matrix is now measurable, for example, with the CSU-CHILL and S-POL radars, it makes it feasible to transform to a circular basis where the separation of shape and orientation effects is more physically direct (e.g., the CDR is independent of canting angle). However, propagation effects such differential propagation phase and differential attenuation have to be corrected for to calculate the ‘intrinsic’ circular observables. We propose that CDR and an orientation parameter (r₄ introduced by Hendry, Antar and McCormick) be used instead of LDR and r_co for possible improvement of hydrometeor classification. The hypothesis is that different classes of particle types can be distinguished due to their differences in shape and fall mode characteristics. Because CDR is independent of canting angle and r₄ is a ‘true’ orientation parameter independent of particle shape, we expect that replacing LDR and r_co by CDR and r₄ will improve hydrometeor type classification. Examples will be shown using CSU-CHILL radar data from the Severe Thunderstorm Electrification and Precipitation Study (STEPS) project.