There too many types of configurations for the Cowan code to handle at once. I ran it once without ng, nh, ni, nk (files with suffix Y) and once including ng, nh, ni, and nk but no higher levels (files with suffix Z). All the observed levels are in suffix Y so those files contain all the *.pos lines. gf2609.pos = gf2609y.pos. Lines that go to ng, nh, ni, and nk are all predicted and have been separated out of gf2609z.lines and combined with gf2609y.lines to form gf2609.lines.