I'd like to thank all of you for replying to my post. You helped a lot. Now I'm trying to find access to S-PLUS software somewhere in my country... Anyway, the summary of all answers can be found in three e-mails I pasted here: >From Joël Mossong <[log in to unmask]> > Dear Srdjan, > > Not sure if this is of much help, but I remember there is a good book by B. > Silverman at the University of Bristol on density estimation using kernels. > That might give you a clue on how to do it, respectively software for it. > > Best wishes, > Joël Mossong >From Adrian Bowman <[log in to unmask]> > If you have access to S-Plus then you might find it useful. > Adelchi Azzalini and I have developed a smoothing library > for S-Plus and this contains a 2-d density contouring procedure > which you might find helpful. The library is available > at www.stats.gla.ac.uk/~adrian/sm > > Best wishes, > > Adrian Bowman >From Noritoshi Tamura <[log in to unmask]> > Greetings, > > SAS JMP software has that capability. I sometime use it . > If you can use S-PLUS software, following information may help you to find a > solution. > > Codes kde2d in directory S of Statlib have S-PLUS's functions of that . > (Pls. see. "Smoothing Methods in Statistics by" Jeffrey S. Simonoff , > http://www.stern.nyu.edu/^jsimonof/SmoothMeth ) > > > Noritoshi Tamura > > Financial Engineering Group, Inc. Thanks again and bye! srdjan %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%