396 MEMOIES OF THE IfATIOXAL ACADEMY OF SCIENCES. 



clamp, C, on tlie earriajje. In the chroiioscope the clockwork is started by pulling a string, and 

 unless the string be pulled with a given force and to a given extent the clockwork is apt not to 

 start properly. We have replaced the string with a bar attached to a telegraphic key, and when 

 the key is tapped the clockwork is properly started. We have replaced the glass bell supplied 

 by the makers with a wooden house. This keeps out dust, deadens the noise, and need not be 

 removed on winding. Lastly, we have rewound the electro-magnets with coarser wire, which 

 greatly reduces the latent time of magnetization and demagnetization. The chronoscope runs 

 one minute only, and must consequently be stojjped after each experiment. A clockwork running 

 a longer time, say one hour, would be much more convenient, and could be regulated more 

 exactly. It would also be convenient if the hands could be sprung back to zero as in an ordinary 

 stop watch. 



The value of the chronoscope consists in the application of an electro-magnet. The hands 

 recording the time are not in connection with the clockwork, and do not move when it is set in 

 motion; but when an electric current is sent through the coil of the electro-magnet the armature 

 is attracted, a system of levers throws the hands into connection with the clocliwork, and tliey 

 are set in motion. Then when the current is broken a spring draws back the armature and the 

 hands stand still.' The distance the hands have moved is read from the two disks, time being 

 recorded to one-thousandth of a second, and the hands returning to their original position every 

 ten seconds. 



It is evident that short times can thus be measured with great convenience. It is easy to 

 close or break an electric current when events occur, and the time is found by subtracting the 

 position of the hands before their motion began from that after it ended. This method of measuring 

 the time of an event is, however, subject to a considerable error, winch was first corrected by one 

 of us. If the strength of the current in relation to that of the spring be too great the recorded 

 times are too long and conversely. This may be conveniently illustrated by a diagram (fig. 2). 



AC D B 



\— 1 



ex D Y 



^. I 1 ' 



C X' D Y' 

 I 1— 1 1 



Flci. 2. 



Let the time be represented by the line A B, and the interval the current is closed (which is the 

 duration of the process) by the line C I). After the current is closed there is a latent period 

 before the armature is attracted and the hands are started, and after the current is broken there 

 is a second latent period before tlie magnetization disappears sufficiently for the spring to draw 

 away the armature. Supposing the tension of the spring to i-emain constant, when a strong current 

 is used the magnetization occurs quickly, say in the time G X, and the demagnetization slowly 

 and the time recorded by the hands A' I'is longer than the real time of the event C B. On the 

 other hand, when a weak current is used the magnetism requires a comparatively long time, say 

 G A'', whereas the demagnetization occurs more quickty, say in D P, and the recorded time is 

 shorter than the real time. This is not a mere matter of theory. With the chronoscope used by 

 us, the real time being 100 o (one-tenth second, (r=0.001 second), the recorded time may have au 

 error greater than 50(7. The discordant results obtained by difii'ereut observers and the large 

 variation in the time of reaction is probably in many cases due to neglect of this factor. With 

 a variable battery (such as Grenet, which is or was formerly supplied by the makers of the 



■ A second electro-magnet makes it possible to reverse this process and measure the time a current has been, 

 broken. 



