4 DISCUSSION OF THE HORIZONTAL COMPONENT 



necessarily from the form and position of the observatory. The effect of the small 

 vertical force bar at first used, upon the bifilar was quite insensible, and that of 

 the declinometer bar affected the value of the scale but slightly, the effect of 

 both instruments changing the value of the scale divisions only in the ratio of 1 to 

 0.9956. 



A thermometer, by Francis, of Philadelphia, divided to half degrees of Fahren- 

 heit's scale, and easily read to tenths, was placed in the box of the horizontal force 

 magnetometer and as near as practicable to the bar. 



After the bifilar was set up, a motion commenced in the direction indicating 

 decrease of force ; it was progressive though not steadily so. After a time an extra 

 scale was required on occasions of auroral, or other disturbances, and finally the 

 ordinary readings were upon this extra scale. On the occasion of the change of 

 the vertical force magnetometer, in January, 1841, by the substitution of Saxton's 

 balance magnetometer for Lloyd's, the magnetism of the horizontal force bar was 

 examined and found to have sensibly decreased ; its force amounted to 0.9601 of 

 its original force, in May, 1840. The experiments were made by means of deflec- 

 tions with a subsidiary declinometer bar, the only means then available. A further 

 experiment of the loss of force was made in June, 1841, when the instrument was 

 accidentally disturbed by one of the observers. The loss of magnetism then found, 

 by means of a new determination of the angle Z, was 0.0314 of its amount in 

 January, 1841. To ascertain the change of magnetism of the bars of the mag- 

 netometers, vibrations were also made use of, but they led to no satisfactory result. 

 The progressive change of the scale readings from the change of the horizontal 

 force and loss of magnetism of the bar, will be investigated further on. 



The observations, between June, 1840, and September, 1843, were made bi-hotirly, 

 and from October, 1843, to the close of the series, hourly. The series extending 

 over five years is not quite continuous; no observations were made on eleven 

 days in January, 1841, on the occasion of the introduction of a new vertical force 

 magnetometer, and the consequent necessity of readjusting the instruments ; in 

 January, February, and March, 1843, the work was reduced to but a single reading 

 a day, by circumstances elsewhere stated ; there are also some minor disturbances 

 at other times when the difference in the readings, however, were ascertained and 

 allowed for. Full statements bearing on the continuity of the series will be given 

 in subsequent pages. 



The reduction proper, necessarily commences with the operation of bringing all 

 the readings to the same standard temperature, to render them comparable among 

 themselves. 



Correction of the Readings of the Bifilar Magnetometer for Changes of Temperature. 



The care bestowed on the experiments to ascertain the effect of the temperature 

 on the instrument, and the perseverance with which they were carried out were 

 not rewarded with a corresponding degree of agreement in the results obtained, by 

 the various processes employed. This it will be recollected was also the case at 

 other observatories. The subject of the co-efficient of temperature for the bifilar 

 magnet is fully treated in the preface to the three volumes containing the record. 



