ABSTRACT
Surely it is true that there is a graded increase of the death rate from cancer of the lung with increase of amount of smoking but equally surely there is an increase of death rate with an increased amount of smoking. If we take the higher death rate from cancer of the lung among the smokers as evidence that smoking causes such cancer then we must also accept that smoking causes cancer of the stomach, of the bladder, of the prostate, for all of which the death rates are higher among the smokers. If smoking were shown to be related to lung cancer in this entirely specific manner then one might feel inclined to support the hypothesis. Consequently, the theory ceases to be specific and becomes as general as to be almost meaningless. Smoking by losing its specific affinity with lung cancer ceases to be an intelligible cause of lung cancer either.
