INTRODUCTIONIn standard normal theory, the statistical procedures are developed under one of the classical assumptions that the underlying parent populatlons are normal in form. The basic reason for this is the spherical symmetry which characterizes normal distribution and the fact that excellent theoretical results can be obtained under normality. In view of the relative abundance of theoretical results in the normal case, there is undoubtedly a temptation to regard any distribution as normal, unless proven otherwise, and to use standard statistical procedures whenever possible. However, experience with real life data reveals that parent populations ocurring in many substantive fields do not, generally, behave in a normal fashion. This fact was recognized long ago by Kepteyn (1903, 1916), Karl Pearson (1905, 1906a and Student (1908 a, b), among others. In this context, the following hyperbolic remarks of Geary (1947b) strike