MEMUlllS OF THE NAT10^'AL ACADEMY OF SCIENCES. 



395 



The point of ap])lic:iti()ii of the stiiuuUis on the body affects tlie U'ngth of the reaction, and 

 this is the problem \yliicli we haA'e more especially attempted to stndy. If the cerebral reflex and 

 motor processes remain the; sanu;, the difference in the time of reaction may be used to measure 

 the velocity of the sensory impulse in the nerve and sjjinal cord. The chief difficulty we have 

 met is not the variable error in the cerebral reflex, but the fa(;t that the same ])hysi('al stinuilus 

 applied at different parts of the body produces physiological effects varying in intensity and 

 cerebral reflexes A-arying- in facility. 



The length of the reaction is also affected by the muscles used, and by using organs at 

 varying distances from the brain (e. g., hand or foot) the velocity of the motor impulse may be 

 studied. The duration of the reaction is the same for the right and left hands, but is shorter 

 when the stimulus is applied to the reacting hand than when applied to the other. In this case, 

 the cerebral reflex is shorter, because it is natural to draw away the hand from a stinuilus which may 

 be hurtful (e. g., a hot surface). The reaction is shorter and more regular when the hand releases 

 a key than when it presses a key — this doubtless because the innervation can not be prepared so 

 thoroughly in advance in the latter case, lest it be discharged prematurely. This is a point which 

 has not been duly regarded by experimenters and has increased the irregularity of some 

 investigations. 



The factors which we have been considering chiefly concern the nature of the cerebral reflex. 

 We, ourselves, believe that the time of transmission of the impulse in the nerve must he far 

 more constant than might be supposed from the discordant results of former investigations. We 

 shall show that the mean variation (and probable error) of a reaction-time may be as sumll as 

 one-thirtieth of the time, and this small variation is doubtless due to changes in the nature 

 of the cerebral processes rather than to alteration in the rate of transmission in the motor and 

 sensory nerves. 



In conclusion, it may be noted that the experiments here described were begun (in 1889) in the 

 psychological laboratory of the University of Pennsylvania and completed (in 18;»3) at Columbia 

 College. The observations were made and the records taken by C. and D. (the writers), and J. 

 (Mrs. J. McKeen Cafctell). The larger part of the new apparatus was secured through au 

 appropriation frcmi the Bache fund of the National Academy of Sciences and is jireserved by 

 the Academj'. 



Part I. — Apparatus and Methods. 



In order to measure a reaction-time at least three instruments are required — one to give 

 the stimulus and record the instant at which it is given, one to record the instant at which a 

 movement is made, and one to nu-asure the inter- 

 vening time. For measuring time we used the 

 Ilipp modification of the Wheatstone electric 

 chronoscope and a new instrument, which we 

 may call a gravity chronometer.' 



The electric chronoscope has been described 

 in various places. It is much more convenient 

 than any chronographi(' method in which a tuning 

 fork writes on a moving surface, and when 

 l)roperly regulated it is fully as accurate. The 

 chronoscope is a clockwork moved by a weight 

 and regulated by a vibrating rod. The rod 

 vibrates (say) five hundred times in a second, 



and at each vibration the tooth of a wheel is allowed to pass, as in the escapement of a clock. 

 The details of the clockwork and method of regulation are not essential, and can be improved. 

 The rod and wheel are shown in fig. 1. The rod is adjusted by the screws s .s, the carriage of 

 the rod moving on the axle A. When the rod vibrates, however, the screws s s become loosened 

 and the rod retreats from the wheel. We have corrected this fault to a large extent by placing a 



Fig. 1. 



' The flectric chronoscopo is made by Peyer & Fararger, Neucbatel. The friavity clirouometer was made by 

 D. G. Brown, Camden, N. J., who also made alterations for us iu tlie electric cbiouoscope. 



