REPORT OF THE SECRETARY. 121 



as great as ten or even twenty seconds. Thus it became customary in 

 describing the sensitiveness of a galvanometer to refer its sensitiveness to 

 a time of single swing of ten seconds. Within the past decade the gal- 

 vanometer needle systems of highest sensitiveness have become relatively 

 microscopic in size and now frequently weigh no more than one or two 

 thousandths of a (/rani (tiro to four millionths of a pound). These systems 

 are often far more sensitive with a time of single swing of only one or two 

 seconds than the best galvanometers of twenty years ago at a time of single 

 swing of twenty seconds. With a needle system practically undamped 

 either by air resistance or induction currents the sensitiveness is propor- 

 tional to the square of the time of swing, so that the sensitiveness of a 

 needle system at ten seconds single swing would on this basis be a hun- 

 dred times that which it would have at a one-second swing. Thus it arises 

 that the computed sensitiveness of these light systems runs perhaps thou- 

 sands of times as great as that of the systems of twenty years ago. But 

 it must not be forgotten that owing to the increased disturbance from 

 mechanical jarring and to the extreme potency of air resistance with these 

 light systems they can not in general be usefully employed at even half 

 so long a time of single swing as ten seconds; and in the second place, if it 

 were indeed possible to use them at a ten-second swing, it would be found 

 that the sensitiveness was perhaps not more than ten instead of a hun- 

 dred fold greater than at one second. Thus comparisons of sensitiveness 

 bused on a ten-second single swing are entirely unfair to the older instru- 

 ments, which could lie and were employed at the time of swing used as 

 the basis of comparison, and hence had a working sensitiveness far more 

 nearly comparable with that of the present day than their computed sensi- 

 tiveness would indicate. In consequence of this unfairness it has recently 

 become common to speak of the sensitiveness at ten seconds <b>nl>l>' swing, 

 a condition at which galvanometers are now sometimes actually used. At 

 this Observatory this change of the basis of comparison has not heretofore 

 been adopted. It must not be inferred from what has been said that the 

 advance made in the last twenty years in the construction of galvanom- 

 eters is belittled, for the reduction in the time of swing for the same 

 degree of sensitiveness is a most valuable saving in time and chances 

 of error, and for automatic recording, as in holographic work, is wholly 

 indispensable. 



In the past two years the design of galvanometer needles has been a sub- 

 ject of much investigation both experimental and theoretical at this 

 observatory, and it is believed that the results arrived at mark practically 

 the limit of probable progress in the way of obtaining sensitiveness at a 

 given short time of swing of a needle system. By tins I mean that it is 

 improbable that a galvanometer can ever be constructed of a given resist- 

 ance which when employed at one second time of a single swing shall give 

 very appreciably greater deflections for given currents than will such a 

 galvanometer as can be constructed with the aid of the knowledge now 

 attained here. In other words, the time for increase of computed sensi- 

 tiveness by tens and hundreds of times witli each newly constructed 

 instrument has passed away. In what has been said I do not wish to 

 claim peculiar advantages for ourgalvanometer, fori understand that both 

 in this country and abroad practically the same results, as regards com- 



