Identification of the true distribution of drop sizes

And adaptive approach has been used to identify the best candidate for the true size distribution of the microcapsules. The underlying procedure is based on linear combinations of the sample order statistics namely, Un and Vn. Various distributions are characterized in terms of the expected values of Un and Vn in a two dimensional plane. Each of the locations and scale family of distributions are represented by a single point, and the distributions having a shape parameter are represented by a curve. For a given sample of size n, the point as an estimate of the true point is determined on a chart of various distributions. The nearest distribution to this point is determined to be the best candidate for the true distribution. It has been shown for most of the cases that the microcapsule data follow a three parameter Weibull distribution, while emulsion data follow Beta distribution.