Introduction
The conventional method of interpretation of grain size data, popular among sedimentologists since the late 1950s, is based on the method of log(ø) probability plots. Grain size distributions of sediment populations transported in nature generally show up as segmented curves when the cumulative frequencies are plotted against logarithms of grain sizes. Phi values of grain sizes (ø=−log2D, where D is grain diameter in mm [Krumbein, 1936]) are generally used for this purpose. A linear phi-probability plot being indicative of a lognormal distribution, the segmented cumulative curves are interpreted as mixtures of several (truncated) lognormal grain size distributions, each corresponding to a particular mode of transportation. It is further assumed that the segments having coarse, medium, and fine modal sizes were transported as bed-, saltation, and suspension loads, respectively. This method of interpretation has developed out of the studies by Inman (1949), Sindowski (1957), Moss (1963), Spencer (1963), and Visher (1965, 1969). The history of the development of the ideas leading to this popular technique of grain size interpretation has been summarized by Visher (1969).
Of late, the validity of this method of interpretation has raised questions in some minds. Cumulative grain size distributions of samples collected wholly from bed or suspension loads, for example, have been shown to plot as segmented curves on probability paper (Sengupta, 1975a,b). It has also been recognized that one-to-one correlation between modes of sediment transportation and segments of log probability plots are based on simplistic assumptions (Komar, 1986).