SIMPLEST NEURAL ARCS. 47 



one apparently relevant difference. On October 8, Subject II remarked 

 that he had "spent all day at the microscope and was tired " 



A second obvious difference in the two days is shown in the course 

 of succeeding periods. On the normal day, succeeding periods after 

 the "normal of the day" show a tendency toward increase in the height 

 of contraction and a reduction of its latency. On the alcohol day, on 

 the contrary, succeeding periods after the "normal of the day" show 

 a gradual increase of latency and a rapid fall in the height of contrac- 

 tion. This change begins within 20 minutes after the ingestion of 

 alcohol and lasts about 90 minutes. At 9 h 3 m , the effect of the 30 gm. 

 hammer had almost disappeared. The substitution of a 50 gm. 

 hammer showed a continual fall of height up to about 90 minutes after 

 the ingestion of alcohol and a slight subsequent recovery. If the data 

 of September 23 stood alone, one could interpret them only as an 

 evidence of the depressing effect of alcohol on the knee-jerk. Taken 

 in connection with the normal record of October 8, the question arises 

 whether the changes on September 23 are not really due to an acci- 

 dental initial extreme excitability and whether the opposite tendency 

 on the normal day is not due to an initial abnormal subexcitability. 

 Subsequent records imply that both of these hypotheses are partially 

 true. 



It is obvious that the least valid measure of the effect of alcohol on 

 the patellar reflex of Subject II would be the difference in the average 

 values of the two days. That would be significant only if they started 

 at the same level. The most significant data are given by the course 

 of the process in succeeding periods after the respective normals of the 

 day, with alcohol and without. If the average of all our cases shows a 

 predominant change in the relation of subsequent measurements to 

 the normals of the day on alcohol days, the direction of that change 

 must be taken as the direction of the probable effect of alcohol. But 

 only if related processes show similar tendencies can we regard this 

 evidence as conclusive. 



All our knee-jerk data are exhibited on this plan in table 2. Each 

 value entered under the appropriate column shows the algebraic dif- 

 ference between the measurements of the first period, or "normal of 

 the day" and each of the succeeding periods of the day. For example, 

 + 5 entered opposite 1 4, under Subject II, October 8, R', shows that 

 on that date the latency of the knee-jerk was 0.005" less in the fourth 

 series than in the first of the same day. 



In the measurements of the patellar reflex, it proved impracticable 

 to follow the usual plan of securing complete sets of comparable data 

 after both doses of alcohol. The extent of the muscle contraction was 

 reduced enormously even by the 30 c.c. dose. In many cases the 

 curves were so low that the latency could not be satisfactorily measured 

 when the action of the alcohol was at its maximum. In most cases 



