langley] personal equation in transit observations 227 



wires of the tally, it will be seen at the second illumination half way 

 between the second and third wires, and so on. 



We have thus arranged that flashes shall occur at recorded in- 

 stants, without limit as to number, and at intervals such that if 

 one, by chance, occurs when the star is bisected by a wire, it will 

 be seen bisected by all the wires in succession. It now remains to 

 arrange that this bisection shall actually occur. This is accom- 

 plished by providing an independent control of the position of the 

 contact points under the clockwork, so that with a cord in the hand 

 of the observer he can cause the series of contacts to occur sooner 

 or later, without altering the interval regulated by the pendulum. 

 Thus if the flash comes a little too early for bisection of the star 

 image by the first wire of the tally, a slight adjustment is made, 

 delaying the succeeding flashes, and with a little practice one or 

 two adjustments suffice to secure bisection. To add the weight 

 of independent observations, several displacements and readjust- 

 ments may be made during the passage of the star over the tally. 



As the times of all the flashes are recorded automatically upon the 

 chronograph, an independent signal is made by the observer to 

 mark each flash which has revealed a satisfactory bisection. 



I have thus far given the device substantially as described in 

 the early paper. To test the value of the method under circum- 

 stances admitting of distinguishing error, an artificial star was 

 arranged to move by an accurate clockwork at about the apparent 

 rate of an actual equatorial star. The shaft which carried the 

 screw by means of which the artificial star was moved, had upon 

 it an arm provided with an adjustable point which instantaneously 

 broke an electrical circuit at each revolution. By careful adjust- 

 ment, zuith the artificial star stationary, it was arranged that exact 

 bisection took place at the middle wire of the tally, and with the 

 point in position to break contact. Thus the artificial star was 

 caused to record its own time of transit upon the chronograph 

 wholly without personal ecjuation of time. 



A transit instrument of 2^4 inches aperture and 48 inches focal 

 length was kindly loaned me by the Superintendent of the United 

 States Coast and Geodetic Survey, for the purpose of making an 

 experimental test of the personal equation machine. 



In reconsidering the theory of the method a possible cause of 

 failure appeared. The observer sees the star continually moving 

 across the limited field of view, and thus his mind may uncon- 

 sciously be influenced by the perception of motion, just as it un- 

 doubtedly is in observations by the usual method. In order to 



