MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 405 



This table fdiitains the result ot more reaetidiis than have been j)iiblislie(l in the many 

 researches hitherto made on the subjeet. The probable eri'or of a single experiment is also 

 much smaller. When the probable error of each measurement is live times as great as in these 

 experiments, as has often been the case, it is evidently necessary to make twenty-five times as 

 many measurements in order to secure a result equally valid. We conse(|uently believe that 

 these exi^erimeuts supersede those hitlierto nuidc. 



We liud that the reaction -time for au electric stimulus a]i])lied to the ui)per arm was 140.6 

 for D and 113.1 a for C, about one-seventh and one ninth second, respectively. The cause of this 

 liersonal difference remains obscure. It must either be due to differences in the nature of the 

 l)rocess or to differences in the sensitiveness of the parts of the nervous system concerned. The 

 maximum and minimum tinu's in the sets differed by 34.7 a for D, and by 11..'! a for C. The differ- 

 ences are not due to the variable error, but to real differences in the con<lition of the nervous 

 system. These are obscure, but the longer times in the case of D were obtained at a time when 

 the nervous system was in a less efficient condition. 



When tlie stimulus was applied on the lower arm or on the leg the reaction time was longer. 

 The excess of time when it was ai)plied on the lower arm was 14.2 a for I) and 6.1 a for C. The 

 maximum and niininuim times difl'ered by 22.3 a for D and 5 for C. These considerable differences 

 are not due to the variable error, but to real differences in the process. The lower point was 30 

 cm. farther from the brain tlian the upper point, and if we can assume the difference in time to be 

 due to the difference in the length of the nerve traveled, we shall have a velocity of the impulse 

 in the median nerve of 21.1 meters per second for D and 49.5 meters jier second for C. The 

 velocities in the sensory fibers of the ])osterior tibial nerve ■\\ould be for I) 31.1 meters per second 

 and for 64.9 meters per second. These times would come within the limits of those oljtained by 

 others, but we are not prepared to accept them as valid. The differences in the times of reaction 

 are undoubtedly correct, and with a very small probable error. The variable error of reactions 

 has been climiiiated, and the times give the reaction-times on objectively equal stimuli applied to 

 the four points. But in the first place it does not seem likely that the velocity of the nervous 

 impulse in the plain nerve should differ so greatly in the two ob.servcrs. We can understand that 

 the entire complex iirocess of reaction might be 2.j per cent longer in one case, as this may be 

 due to less complete coordination in the brain centers, but there seems to be no reason why the 

 velocity in the plain nerve should be 140 per cent slower. This would require us to conclude that 

 the central reflex was actually shorter in the case of D. In the second ])lace the probable error 

 in the case of the velocity of the impulse is far greater than it should be from the variations in 

 the reactions. 



We are consequently compelled to conclude that the differences in the reaction-times are due 

 to differences in the cerebral processes, and not merely to the length of the nerve traveled. 

 The times for the points nearer the brain may be shorter because the physiological effects of the 

 shock were greater (as was iu fact the case) or because the fibers from the upper points lead to a 

 more rapid transference in the brain. The difference iu the two observers and in the same observer 

 at different times are not due to variation in the velocity of the nervous impulse, but to difler- 

 ences in the nature of the cerebral reflex. For example, C's reaction time being shorter and more 

 regular than D's, we may conclude that it is more automatic and less influenced by changing 

 conditions. The greater intensity and massivene.ss of the shock on the upper points would 

 consequently shorten the reaction-time less for C than for D, and the results iu the table are 

 thus explained. This would also explain the larger variation iu the sets in the case of 1). This 

 variation may be partly attributed to differences in the relative effect of the shock on the two 

 points at different times, and may be partly due to cerebral changes. These considerations led 

 to further experiments by which they were fully confirmed. 



When the shock was applied to the leg in one case and to the arm in tiie other tlie impulse in 

 the former case had in addition to travel through the spinal cord from the lumbar to the brachial 

 plexus and the times are considerably longer. The difference in reaction-time with the hand when 

 the' shock was applied to the upper leg and upper arm, respectively, was for D 25 a for C 26.6 0. 

 .When the shock was applied to the lower leg and arm, resi)ectively, the diflerences were for D 

 26.9 ff, and for C 28.1 o. These times agree very well, better, perhaps, than the probable error would 



