Introduction. 9 



meriting. This should not be construed as signifying in any way or 

 manner a doubt concerning the experiments and findings of Dodge and 

 Benedict. That these deserve the confidence of the reader has been 

 proved by a rechecking of the experiments from original protocols and 

 in many cases a recalculation of the results with very minor corrections, 

 as indicated in the following pages. The new set of experiments on 

 Subject VI has shown so many agreements with the previous findings 

 for this subject as to convince us of the worth of this method of attack- 

 ing the problem. Furthermore, the comparison experiments have 

 established certain individual differences as to the effect of alcohol upon 

 this subject which could not otherwise well have been proved. 



There is ample precedent for the publication of experimental results 

 obtained on only one subject; in fact, most alcohol researches have em- 

 ployed a very limited number of subjects. Ach, 1 Frankfurter, 2 

 Furer, 3 Hellsten, 4 Lombard, 5 Mayer, 6 Scheffer, 7 Totterman, 8 and Vogt, 9 

 as well as others, report experiments performed on one individual. The 

 studies mentioned are among those in which the author has in each 

 instance experimented on himself. This method of experimental self- 

 observation, while having certain advantages, can not usually be 

 depended upon to produce the objective results which are so much to 

 be desired in this field. 



Ideally, alcohol research should be conducted in such a way that 

 adequate data would be recorded and reported on each one of a large 

 number of subjects who are chosen in accordance with certain classifi- 

 cations. The results, then, for the individual and for the group would 

 be adapted to statistical treatment. The obstacles which confront the 

 single investigator who would complete such a research on human sub- 

 jects, while not insurmountable, are nevertheless such as to make it a 

 labor of years. In the meantime other investigators, who are working 

 on the same or related problems, should have some of the data made 

 available for their use and criticism. Logically one of the first steps 

 toward the broader solution of the problem is to discover how large a 

 variety and how many measurements of each process must be taken 

 on the individual in order to obtain a satisfactory sample of his perform- 

 ance under alcohol. The data for the individual must be subjected to 

 statistical treatment in so far as practicable and the experiments re- 

 peated to discover agreements or disagreements between successive sets 

 of results. Accordingly, while reports on single subjects may be con- 



1 Ach, Kraepelin's Psychol. Arbeit., 1901, 3, p. 203. 



2 Frankfurther, Kraepelin's Psychol. Arbeit., 1914, 6, p. 419. 



3 Furer, Ber. ub. d. V. intern. Kongr. z. Bekampf. d. Missbr. geist. Getrankes zu Basel, 1896, 

 p. 355. 



4 Hellsten, Skand. Arch. f. Physiol., 1904, 16, p. 139. 



5 Lombard, Journ. Phy«iol., 1893, 13, p. 1. 



6 Mayer, Kraepelin's Psychol. Arbeit., 1901, 3, 535. 



7 Scheffer, Arch. f. exp. Pathol, u. Pharmakol., 1900, 44, p. 24. 



8 Totterman, Finska Lakaresallskapets Handlingar, 1916, 58, p. 1527. 



9 Vogt, Norsk Magazin for Leagevidenskaber, 1910, 8, p. 605. 



