COMPLEX NEURAL ARCS. 101 



breaker closes again, and the marker returns to its normal position. 

 The armature of the exposure apparatus, however, is so far from its 

 magnet that it remains unaffected by the closure of the circuit and the 

 exposure is continuous. When the subject reacts by speaking the 

 word, the circuit of the electric marker is broken a second time by the 

 voice key. This second movement of the marker indicates the moment 

 of reaction. Since both the exposure and the reaction are recorded by 

 breaking the same electric circuit, and since both events are recorded 

 by movements of the same writing-point, the alignment of the marker 

 and its latency do not affect the records. 



Standard instructions to the subject. (1) Hold the voice key to the 

 mouth, pressing it firmly against the upper lip ; (2) speak the words as 

 soon as possible after you see them; (3) if you misread or mispronounce 

 a word, speak it correctly as soon as possible. 



RECORDS. 



A typical word-reaction record is reproduced in figure 22. The record 

 reads from right to left . The extreme right-hand breaks in the horizontal 

 lines indicate the moment of contact between the offset of the kymo- 

 graph shaft and the circuit-breaker which was in series with the marker 

 and the exposure apparatus. These breaks are in approximate vertical 

 alignment. Since the kymograph drum revolves completely in 5", and 

 the stimulus follows at intervals of 10", each alternate break is insig- 

 nificant and is not followed by a reaction, because the exposure appa- 

 ratus was arbitrarily prevented from falling by the operator. The 

 second elevation in the horizontal lines of the significant records is 

 the reaction break. The character of the reaction record varies with 

 the word. The continuous record furnishes its own distribution curve. 

 Measuring the distance between the stimulus and the reaction breaks 

 gives the reaction-time. One millimeter along the base-line equals lOcr 

 (0.01 "). All records which are not marked at the time of taking, as 

 defective in technique or in response, are included in the following 

 results. In this, as in other measurements, we deemed it inexpedient 

 to eliminate any records on the basis of probable error, unless the evi- 

 dence of inadequacy was given, independent of results. 



RESULTS. 



Table 10 gives the data for word-reaction experiments. Results of 

 the normal days are entered on the left, alcohol days on the right. 

 In the first column of each section are entered the designation of the 

 subject, the date of the experiment, and the number of the periods. 

 In the next column are entered the averages of the 24 reactions of each 

 period. Their mean variations are given in the column headed Mean 

 variation. Under the heading Difference are given the deviations of 

 subsequent periods from the first, according to the formula D = 1-2, 

 1-3, 1-4, etc., and also the mean variations of these differences. 



