A general description and analysis of the hyperbolic tangent distribut
ion family are presented. Analytical formulae are given for both the c
umulative and frequency functions of the distribution, as well as for
the ordinary moments of whole order. A linear transformation of the si
ze coordinate axis is applied by means of which the distribution funct
ion can be located appropriately over the interval between the observe
d smallest and largest sizes of particles during the fitting procedure
. Including the parameters of this transformation, the distribution fa
mily possesses four parameters, which allows for effective fitting to
experimental size distribution data. A two-level fitting procedure has
been developed which was tested by using simulated noisy data. A numb
er of distribution functions (log-normal, Rosin-Rammler, and beta dist
ributions) may also be described well by hyperbolic tangent distributi
on functions, so that it provides a convenient method for comparing me
asurement data described quantitatively in different ways. (C) 1998 El
sevier Science S.A. All rights reserved.