Program: maxdip Purpose: To estimate rho (4Nr) and gene conversion rates using random samples with two-site unphased diploid information. The estimate is obtained by a maximum composite likelihood method based on pairs of sites, described in Hudson, R.R. 2001. Two-Locus Sampling Distributions and Their Application Genetics 159: 1805-1817, though gene conversion is not included there. The program accomodates missing data, and can use information on the ancestral state of a site, if such information is available. An infinite-sites model is assumed and sites with more that two-alleles are ignored. To use the program, one must have a file of two-site (haploid) configuration probabilities for an appropriate range of recombination rates and for samples as big or bigger than the data sample. Examples of these configuration files are available at http://home.uchicago.edu/~rhudson1/source/twolocus/tables.html The files at this site are for samples of size 100, 50 and 30 and are named h100rho, h50rho, and h30rho, respectively. Also, at this site there is a program (makenewh.c) for generating files for smaller samples from a file for a larger sample. Also available is a program to generate these tables from scratch for any sample size. For large samples the program must be run for a long time. Usage: maxdip numfiles likefile1 likefile2 ... likefile_numfiles rho_start maxfactor minfactor fmin fmax fnum [-g] conlen [-p] R4it.pairs will produce a file suitable for use with maxdip. Then: maxdip 1 h30rho .001 50. .01 0.0 0.0 1 400