Twelve methods of fitting the two-parameter gamma distribution to unit hydrographs are evaluated using experimental data. Five of them are the methods of moments (MOM), cumulants (MOC), maximum likelihood estimation (MLE), least squares (MOLS), and principle of maximum entropy (POME). The remaining seven are based on specification of different point and planar boundary conditions. It is found that of the later seven methods, only one is satisfactory; the other six are inaccurate and should be abandoned for purposes of achieving a mathematical fit. The first five methods are reliable, easier and efficient, and should be preferred.