Would a better understanding of the anatomy and improvement of the method make it possible to reduce the percentage of failures sufficiently to earn selective arteriography of bronchial and intercostal arteries a place among the accepted methods of selective arteriographic investigation?