122 . RKPORT OF THE SECRETARY. 



jiuted sensitiveness, have recently been reached by several physicists 

 independently, which strengthens the view that little further advance in 

 this direction is likely. 



But the useful or working sensitiveness of a galvanometer is another 

 matter, and by the system of support and magnetic shielding described in 

 my last year's report great advantage has been gained in this already, and 

 still better results are hoped for by still other improvements. Let me 

 clearly indicate how progress in wf)rking sensitiveness may be consistent 

 with a standstill in computed sensitiveness. The spot of light reflected 

 from the mirror of the galvanometer needle, which should be quiet when 

 no current is being observed, is always making slight excursions upon the 

 scale, and these fluctuations prevent readings of current deflections to be 

 made of less than a (certain minimum amplitude, for they then become 

 indistinguisiia])le among the accidental deflections just mentioned. Let 

 us now suppose tliat the average of accidental deflections should be reduced 

 by better elimination of ground tremors and magnetic fluctuations from a 

 millimeter to a tenth of a millimeter on the scale, then it is apparent that 

 ten times the working sensitiveness is attained. But let us suppose that 

 further improvement in these res])ects is found possible. It is hardly 

 jiracticable to read the position of an ordinary spot of light more accu- 

 rately than to the nearest tenth millimeter, so that little progress would 

 directly result, but the time of swing of the needle might be profitably 

 increased. Then, however, the effect of air damping would soon become 

 so prejudicial as to stop advance. 



We are now in position to state generally the methods employed and 

 the results attained and hoped for here in this matter of increasing the 

 working sensitiveness. The aim of all efforts is to make it possible to read 

 deflections to a tenth of a millimeter on a scale at 3 meters with an actual 

 time of single swing of ten seconds. 



In the first place it has been sought to reduce the mechanical tremors of 

 the galvanometer due to the city traffic; and for this purpose the elaborate 

 pier and suspension system described in my last year's report was con- 

 structed. In the second place it has been atteniitted to reduce the prejudi- 

 cial effects of these and other mechanical disturbing factors which still 

 remain to ]\iv the needle itself. To fully understand what has been planned 

 for this purj)Ose it should be stated that in addition to such mechanical 

 tremors as have already been referred to, it has been found that the sound 

 waves sent out from concussions of various kinds are able to seriously 

 affect the steadiness of the needle. These sound waves can travel into the 

 galvanometer case to jar the neeille despite any system of support, and the 

 only way to avoid them is to exhaust the air in the galvanometer, so that 

 our new cases are of air-tight construction. The exhaustion of the air, in 

 addition to preventing disturbance by sound waves, also viakps the anisi- 

 live)iess nearh/ ])roporlional to llw square of time of Kiviny of (lie needle, so that 

 it is no longer so unjust to use a ten-second time of single swing as the 

 basis of comparison. But in addition to securing exhaustion of the air as 

 a means of reducing mechanical tremors, another device has been found. 

 The experimental and theoretical investigations of needle systems above 

 alluded to have indicated a method of construction by means of which the 

 veiglit of the needle sydem can be largehj Increased wilhoat diininiftliing the com- 



