In the application of Lagrangian Stochastic Models (LSM) for turbulent 
dispersion of passive tracers a closure problem arises in the determination of 
the shape of the probability density function (pdf) of Eulerian velocity. In 
this paper commonly used closure schemes are reviewed and the importance of 
using the maximum missing information (mmi) pdf is outlined. Analysis of 
available data from literature on 3rd and 4th order pdf moments shows the need 
for using non Gaussian pdfs and suggests a possible closure through a 
relationship between skewness (S) and kurtosis (K)