Correct statistical treatment of results of Goss and Harris experiment
s on radiation hybrid mapping is proposed. Using stochastic simulation
a problem of an optimal sample size is investigated. Reasons for a re
petition of experiments with different doses of radiation are discusse
d.