The failure of inactivation was evident because growth was being detected on multiple production runs, these runs were routinely sent back for additional radiation. It should have been seen for what it was: an indication that the margin of safety with the method was not sufficient.