( 241 ) 



The regiilarl)' beul line of mediiiiii (hickiiess repi'cjsciil.s the I'ccorded 

 curve. At the moment t = Ihc constant current is nuide. In the 

 case of an ideal, absolutely accurate analysis we should find two 

 straight lines, one of. which would rise vertically from ^ to i?, while 

 the other from J3 to C would be horizontal. The results of (he actual 

 analysis according- to column 5 of table IX are I'cprcsented by the 

 thick line, while the thin line represents the values of rc^' accord ing 

 to column 3 of this table. 



For the virtual mass m two dilferent values have been used ; for 

 the first 4 thousandths of a second in has been put 0.0567 which 

 is 6 times the amount found in chapter 3. At t = Q o and t = S m, 

 ?/i has no influence, since at these times q may be put = go. Begin- 

 ning with t = 10 o, m has again been reckoned but (his time with 

 a value 0,0187 which is twice the value of chapter 3. 



If a single value is given to m the results become much less 

 satisfactory and the question arises whether the whole analysis nnist 

 not be considered worthless now that it appears to be impossible to 

 assume a constant value for m. 



In rejily to this we remark in the first place that, as will pre- 

 sendy appear, the variation in the value of m is only of account 

 with a great sensitiveness of the galvanometer, i. e. wi(h a feeble 

 tension of the quartz thread. Besides, even with the most sensitive 

 position of the galvanometer still an important part of the analysis 

 can be usefully applied. For practically a curve, such as is obtained 

 e. g. in many electrophysiological investigations, will consist of parts 

 of various curvatures and will always show a number of points for 

 which Q may be put := go or the acceleration may be put = 0, 

 In all these points ))i need not be taken ijito account. Since r can 

 be measured with great accuracy the analysis is in all respects 

 satisfactory here. 



Moreover the analysis can be applied wherever the curvature 

 and at the same time the angle of inclination are not too great — 

 in our case already in all points that are recoixlcd later than 0.004 

 second after the starting of the current — as will appear from table 

 IX and figure 3. For each definite tension of the quartz thread a 

 definite value for m be taken. 



The reason why in general m is represented by another value 

 for different tensions of the quartz thread will be discussed in chapter 

 8. But here we must ask why in can also vary Avhen the sensitiveness 

 of the galvanometer and (ogedier with it the tension of the quar(z 

 thread remain unchanged. For an explanation of this unforeseen 

 and somewhat disappointing phenomenon we ha\'e iji the first 



17 



Proceedings Royal Acad. Amsterdam. Vol. VIII. 



