XVI INTRODUCTION. 



under the telescope wire. By repetition of such measures the angular value may be determined 

 with any required accuracy. The ascertained value is 1'. 00896. 



To obtain the actual variations of the declination the observed changes from the magnetic 

 meridian in scale divisions must be multiplied by the arc value corrected for the ratio of the 

 torsion force of the suspending fibres to the earth's directive force, or by the coefficient 



a (i _j_ 5Y The value of 5 is determined by turning the torsion circle of the suspension 



V F / 



tube through two or more large angles, (for example 90,) and noting the corresponding dif- 

 ferences of scale reading ; then if w equals the mean of the former, and u that of the latter in 

 angular measure, 



H_ u 

 F w u 



The same suspension fibre was used for the horizontal force experiments and subsequent obser- 



TT 



vations of diurnal variation, and the value of the coefficient for each day will be found 



with the series first named, and printed on pages 120-123. 



There were vertical iron bars to the windows within 20 feet, and iron fastenings to doors 

 within half that distance of the instrument, but their influences on the magnet were constant, 

 or, at most, were variable only by differential inductive changes, produced by the earth quan- 

 tities probably too minute for appreciation with our means. 



COMMENTS AND ABSTRACTS. 



Ordinarily, the north end of the declination magnet, then at its westerly limit, begins 

 moving to the eastward at 97&. Om. A. M. mean time, Santiago. This hour varies with the 

 season of the year, being as early as 8ft. Vim. in spring, and as late as 9A. 53m. in winter. 

 Taking an average of the whole year, its maximum of easterly deviation is reached at lh. 45m. 

 p. M., the angle described in the intermediate time being 6'. 88. In summer, the period of the 

 maximum is retarded about 18m., but in winter it is anticipated a like interval, spring being 

 2m. later, and autumn 5m. earlier than the annual mean. The angular fluctuations vary in 

 like manner, being greatest in spring and least in winter, autumn coinciding very nearly with 

 the mean for the year, and summer differing but little from spring. In the first named season 

 the arc described is 8'. 90, and in the second 3'. 96. 



From 1A. 45m. p. M., to about 77i. 05m. p. M., the movement is westerly, and the arc de- 

 scribed 3'. 96. Combining the observations of each season to ascertain the several periods of the 

 latter, autumn coincides most closely with the mean of the year, and spring differs the most 

 widely from it, the mean instant being retarded 5m. in the former, and anticipated 40m. in the 

 latter season, whilst there are delays of 16m. and 18m. in the winter and summer respectively. 

 The angular value of the westerly movement between these Hours is much more uniform than 

 the antecedent easterly oscillation, the largest arc described, 4'. 54, being that of summer, and 

 the smallest, 3'. 61, that of winter. Autumn and spring are 3'. 67 and 4'. 04, respectively. 



Between *lh. 5m. p. M. and about 6A. A. M., the north end of the magnet again moves to the 

 eastward, the arc traversed being small ; but as there were very few observations between mid- 

 night and 6 A. M., both the period and the arc must be regarded as only approximately known. 

 Moreover, an examination of the record shows that the nights are uniformly more subject to 

 abnormal oscillations than the days, and these frequently mask not only the instant of the lesser 

 easterly departure from the mean magnetic meridian, but also the anterior hour of the lesser 

 westerly deviation. 



Finally, between 6A. A. M. and 9A. Om., the bar has rapidly moved through an arc of 4'. 20 

 to its extreme westerly plane. Of the several quarters the angular value of the oscillation in 



