A comment on the ratio of average width to average spacing of nuclear resonances