20 PSYCHOLOGICAL EFFECTS OF ALCOHOL. 



some modifications of the measurable qualities of selected neuro-mus- 

 cular processes occur more regularly or in greater or less degree after 

 the ingestion of alcohol than without it. Giving due weight to our 

 measurements of the normal variations, we can say that the average 

 change in the measurable qualities of the selected processes after the 

 ingestion of alcohol, minus the average change under otherwise similar 

 conditions, but without the ingestion of alcohol, will represent the 

 effects of the dose of alcohol that was administered. Experimental 

 results of this sort have a degree of probability which depends not only 

 on the accuracy of the individual measurements and the similarity of 

 controllable circumstances, but also on the number of experimental 

 instances and on the probability of a really chance distribution of the 

 accidental variations. The sequel will show that for no single subject 

 are the data sufficiently numerous, except in the case of the pulse- 

 records, to give a satisfactory quantitative statement of the individual 

 differences of the effect of alcohol. Our experimental answer to the 

 main question at issue, viz, as to the general direction and amount of 

 change in the various processes consequent to the experimental inges- 

 tion of alcohol, is, we believe, conclusive and adequate. 



In addition to the main experimental precautions, w r e systematically 

 varied the alcohol dose. This was done for the following reasons: In 

 the first place, it is a fact that different doses of some drugs produce 

 quite different physiological effects, amounting even to a change of 

 sign. That this is probably true of alcohol seemed to be indicated in 

 more than one experimental investigation. The existence and con- 

 ditions of such a change in the effect of alcohol, if it really occurs, 

 is a peculiarly important phase of the alcohol problem. In the second 

 place, we felt that no safeguard against mistaking accidental variation 

 for causal relationship is so effective and no evidence is quite so con- 

 vincing as that of concomitant variation in the amount of the alcohol 

 dose and its effects. We believe that the results justify the increased 

 labor, and that in no other way could we have secured the same insight 

 into the vagaries of the commonly observed effects of alcohol. 



NORMAL OR BASAL EXPERIMENTS. 



The fundamental requirements of method which we have already 

 considered demand the largest possible number of measurements of 

 the phenomena under investigation, both with and without alcohol, 

 but under otherwise similar or comparable circumstances. On general 

 logical principles, the number of instances should be approximately 

 equal in both cases. This was arranged for in our routine, by the 

 regular introduction of normal days identical with the alcohol days 

 as far as practicable, with the exception that on normal days no alcohol 

 was administered. Furthermore, even on alcohol days one normal 

 period was given before the dose. Each experimental session may 



