Some clinical inferences from the study of serial laryngeal sections