REPORT OF THE SECRETARY. 93 



the aim last year set forth to carry the proportionahty Ijetween deflection and square 

 of tlie time of swing np ti> lOl^econds single swing will at length be reached. (Ireat 

 steadiness was observed at 5:5. seconds swing. A deflection of 1 millimeter (in the 

 scale then actually corresponded to a current of only ioooo(r()oon(]o amperes in the 

 galvanometer. The resistance of the instrument is only 1.6 ohms. A current of one- 

 tenth this magnitude — or 5X10 " amj)eres — could actually have lieen measured. 

 Further advance is confidently expected. 



Permnal-etjiKtfion iti<irji!iu\ — You have placed at the (Jl)servatory for trial an instru- 

 ment of your own design intended to eliminate the so-called personal equation of 

 individual observers in transit observations. The principle of this instrument con- 

 sists in substituting a judgment of the place where a sudden phenomenon occurred 

 for the time when it occurred. To use an illustration which you have already 

 employed, in case the dark field in which only the star is seen moving were illumi- 

 nated every two seconds by a self-recording flash whicli showed the central wire, 

 and if by pure accident the star was caught in an exact bisection when the flash 

 came — an accident against which the chances are perhaps more than a hundred 

 to one — it is evident that in this rare and improbable event there would be no jier- 

 sonal equation to allow for, if the time of flash within two seconds were noted by 

 the observer. Now, the object of the following mechanism maybe said to be to 

 make this accident hap])en every time. 



This being understood, as first tried with the apparatus which you furnished, the 

 design was to ilUiminate the cross wires of the transit instrument by automatically 

 recorded electric flashes occurring at regular intervals equal to the time elapsing 

 between passages of the star acro.ss successive wires. An adjustal)lemechanism allowed 

 the ol)server to hasten or defer the whole system of flashes until by successive adjust- 

 ments he caught the exact instant when the star was bisected by the dark wire in 

 the instantaneously l)right field. Several such adjustments could l)e made (hiring a 

 single star transit across the numerous wires of the tally, and immediately after each 

 satisfactory l)isection a signal was made on the chronograph by a key in the hand of 

 the observer, so that only such recorded flashes as were thus distinguished were used 

 in determining the time of transit. 



In order to test the value of this instrument an artificial star was caused to move 

 through the field at a rate about equal to that of a real equatorial star. This star 

 was moved by a screw and clockwork of great accuracy, and always through the 

 same portion of the screw. It was also provided that when the star was exactly 

 bisected l)y the middle wire an electrical contact was broken, so that the star recorded 

 its own transit upon the chronograpli. The adjustment of the contact for this 1)isec- 

 tion was made at leisure with the star stationary, and was therefore not su])ject to 

 the personal error of transit observations. Accordingly, after determining all the 

 wire intervals, it was entirely easy to measure the personal etiuation of the observer, 

 whether he used the personal-equation machine or the ordinary method, merely by 

 comparing his observed time of transit with that recorded by the star itself. Fpon 

 trial it was found that the personal equation of the several observers was not wholly 

 l)revented by employing the machine in the manner described; that is, by bisections 

 judged during instantaneous flashes. The writer, for example, observed about G. 1:.* 

 second too early by both methods. 



It seemed probable that the observer was still l^iased in his judgment by watching 

 the march of the star through the field across the very faintly showing wires, which 

 are always to be made out by the stray light of the star and sky, even though there 

 is no illumination by the flash. Accordingly it seemed to promise success to alter 

 the arrangement so that the star would be hidden except at the instants correspond- 

 ing to those when the flash had formerly appeared, and to steadily illuminate the 

 field as is usual in transit instruments. In the test this was easier than in actual 



