24 On the Calibration of the Capillary Electrometer. [Nov. 21. 



constants representing the variations of this ratio for different values 

 of the acting P.Dk It is hardly necessary to point out that this is 

 quite a different thing from my " calibration curve, '^which relates 

 to the variation of the time-relations at different parts of the same 

 capillary. 



With regard to Professor Einthoven's observations, I can only say 

 that I have not met with the phenomenon save in " sticky" electro- 

 meters ; and in such cases minute irregularities are plainly visible on 

 the curves, when the plates are driven fast enough. But the rate I 

 employ for physical investigations, namely, from 1,000 to 1,500 mm. 

 per second, is so many times greater than that used in the examples 

 he gives, viz., 20 to 25 mm., that I am able to detect what would 

 otherwise be invisible. His instruments would appear to be much 

 less rapid and less sensitive than most of mine, but he uses a higher 

 magnifying power for the projection. The actual distance traversed 

 by the meniscus itself in making the excursion, of which the 

 analysis was given in my paper,* was fifteen times as great as that 

 traversed by the meniscus of his electrometers in the eight excur- 

 sions adduced in support of his new theory. But although those 

 excursions are smaller and slower, the curves are much steeper than 

 I should deem advisable, especially as the length measured, according 

 to his plan, is the reciprocal of the ratio to be determined. 



That there must have been either irregularities in the curves, or 

 considerable error in the measurements, is manifest when, instead of 

 taking the average values as he has done, the eight curves are reduced 

 to a common scale, and plotted on the same sheet of paper. One of 

 them especially, No. 205 A, differs from all the rest in form. 



By using the asymptote to the normal curve as a line of reference, 

 Professor Einthoven introduces a source of error. The slightest sticki- 

 ness may cause the meniscus to stop short of the asymptote, or an 

 insignificant amount of overshooting may carry it beyond, and a 

 difference which would appear in the fifth place of decimals in esti- 

 mating an E.M.E. becomes of crucial importance in calculating the 

 time-relations of the movement. It is for this reason that I use the 

 upper limit of the photograph as a reference circle, and reckon the 

 displacement of the meniscus from its zero-position. 



I have, since reading Professor Einthoven's paper, again calibrated, 

 by both the methods described in this communication, the electro- 

 meter in use in the Physiological Laboratory at Oxford. Both cali- 

 brations agree, proving that the instrument shows no trace of the 

 peculiarity mentioned by him. Should I detect it in any future 

 electrometers, I will not fail to communicate the fact to the Society. 



* ' Philosophical Transactions/ vol. 183, A, p. 95. 



