1895.] On the Calibration of the Capillary Electrometer. 23 



The numbers in the fifth column are obtained from those in the 

 fourth column by dividing each of them by 55. The last column 

 needs no explanation. The time spent in taking these photographs, 

 and another set nearer the tip of the tube, was about one hour, and 

 the calibration of the scale readings and measurement of the sub- 

 normals of both sets, was effected in an hour and a half. The entire 

 calibration covered 95 mm. of the projected image of the capillary, 

 but the upper part was completely spoilt by the stickiness of the 

 tube. One advantage of this process is that it enables such defects 

 to be detected by inspection. Each mechanical obstruction leaves a 

 mark on both curves, but most evidently on that which has the 

 smaller subnormal. The effect is equivalent to a loss of time in the 

 excursion. The meniscus is retarded in the act of passing the point, 

 but immediately afterwards resumes its normal velocity. It would 

 be quite possible, if it were worth while, to measure the "work 

 done " in passing a sticky place. 



It should be observed that if the value of the scale-readings is 

 known, the calibration of the subnormal can be effected by means of 

 a single excursion of known P.D., provided that the zero-point is 

 visible on the photograph. For the acting P.D. at any point on the 

 curve can be found by subtracting from the E.M.F. given by the 

 potentiometer, the P.D. corresponding to the rise of the meniscus ; 



and the ratio ac ^ n ff PD- £ Qr a ser j es f equidistant points, gives the 

 subnormal 



calibration-curve of the capillary. This was the method I employed 

 in 1891, but it requires much greater care to get concordant results, 

 and I realised that until some simpler process could be found, few 

 people would care to use this method of determining the value of 

 rapid changes of potential difference with the capillary electro- 

 meter. 



Note. — Professor Einthoven, in a paper published in 1894,* recom- 

 mends that in selecting an instrument, care should be taken that it 

 is of equal sensitiveness throughout the working portion of the tube, 

 and assumes that such a test is a sufficient criterion of its fitness for 

 use. 



This I have not found to be the case. A constant subnormal 

 is always associated with a slight increase of sensitiveness towards the 

 tip of the capillary. 



More recentlyf he has put forward a statement having a very 

 important bearing on this problem. He now finds that in some 

 electrometers the ratio of the velocity of the meniscus to the acting 

 P.D. is not constant, but diminishes as the mercury approaches its 

 position of rest. In such cases he proposes to construct a curve of 



* ' Archiv fur die Gres. Physiologic/ vol. 56, p. 528. 



f ' Archiv fur die G-es. Physiologic,' vol. 60, pp 91— 100> and 101—123. 



