On a branching process in neutron transport theory