Analysis of neuronal soma size distributions.

Neuronal soma size distributions are often very skewed, leading to difficulty in their characterization and in the application of statistics commonly used to describe normally distributed variables. A particular family of curves, called Weibull functions, can be shown to fit cell size data extremely well. These functions not only characterize these skewed distributions, but do so in a way which permits powerful, sensitive comparisons of differences between distributions. Because of the flexibility of the Weibull functions, they may be used to describe a variety of morphological attributes. Several applications of these functions to neuroanatomical data are described using methods which require only graph paper and hand calculator for curve fitting. An extension of the basic curve fitting method permits mixtures of Weibull functions to be used in describing multimodal size histograms.