Using 456 data sets under bank-full conditions obtained from various sources, the geometric relations, derived in part 1 <Singh et al., 2003>, are calibrated and verified using the split sampling approach. The calibration of parameters shows that the change in stream power is not shared equally among hydraulic variables and that the unevenness depends on the boundary conditions to be satisfied by the channel under consideration. The agreement between the observed values of the hydraulic variables and those predicted by the derived relations is close for the verification data set and lends credence to the hypotheses employed in this study.