That the problem of controlling the drinking driver and pedestrian is far from being solved may be due, in part, to the fact that no accurate statistics are available regarding the relationship of alcohol to automobile accidents, as well as to the fact that the diagnosis of drunkenness is, to a great extent, still made through unreliable physical observations.
In a previous paper1 the chemical test for alcohol in body fluids has been shown to be specific and a practical method for confirming a diagnosis of drunkenness, thus aiding in the conviction of drunken drivers.
In the present paper a further study has been made of the subjective and objective symptoms due to a consumption of alcohol correlated with the chemical examinations, and an analysis of 119 automobile accidents involving injury or death to 216 persons has been made in an effort to obtain a closer estimate of the