Greetings -
Gary has posted a new MARK build, which corrects a few issues with calculation of the likeilood for saturated models (in particular, wrt to Pradel models, and models with individual covariates).
You can download it from
http://www.phidot.org/software/mark/download/