I believe the `splancs' package from CRAN has a bivariate K function. For error bars you'll probably have to use Monte Carlo methods. `splancs' has some tools for that.
See also 'spatstat' for Kmulti and Kcross.
Since R doesn't have a univariate K-function in the base or recommended packages I'm wondering which package the original poster got univariate K-function from. I suspect their friendly sysop installed splancs or spatstat.
Baz
______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
