KXl'KIHTION TO I'OINT IJAItUONV, ALASKA. 451 



'To determine the torsion coefficient - 1 '. let = angle through which the magnet was deflected, 



and ,i = angle through which the torsion circle had been turned; then * = , ' ; hence scale valuo 



./ ft a. 



a ( l + -i- ) expressed in minutes of arc. Increasing numbers of scale should correspond to a mo- 

 tion of the iwrth end of the magnet to the east. The scale is numbered from 20 to 80, which numbers 

 are to be read 200 and 800, and thus has a range of 5 on either side of the normal position. Two 

 spare scales, divided on white bristol board, about 15 centimeters long, giving additional extent of 

 2, should be made, and, in case of necessity, fastened to the ends of the reading scale. The 

 vertical cross-thread of the telescope is to be l-<'/>l on the .">00 mark, as reflected from the lixed 

 mirror,* a remark which applies to each of the instruments. The dividing line or narrow space 

 between the fixed and movable mirrors is in the plane of the optical axis of the telescope. Tho 

 instrument is placed under a zinc cover. 



"2. THE HORIZONTAL FORCE OR BIFILAR MAGNETOMETER. 



"Put plane of detorsion in the magnetic meridian, turn torsion circle with weight suspended 

 approximately in plane of meridian, and read circle. Remove weight, suspend magnet, and again 

 read circle, if the same as before the plane of detorsion is in the magnetic a , Ha 

 meridian: if not, repeat the process until the result is satisfactory. It is I 

 recommended to mark out in the observatory the directions of the mag- 

 netic meridian and of the magnetic prime vertical by threads or fine 

 strings stretched from wall to wall. These threads would also aid in the 

 setting of the piers. Let m= reading of torsion circle for plane of de- 

 torsion in the meridian: suspend weight and turn torsion circle to 90+ m; 



turn movable mirror until the middle line or 50 of the scale is bisected, in 

 which position of the telescope the fixed mirror will reflect division 50 (to 

 be read and recorded as before 500). Suspend magnet in place of the 

 weight, turn torsion to wi until middle line of scale is again bisected. 



then >n^' l (fl0 + ?H)=.:. (See annexed diagram, where =DO.) Let H = horizontal component 

 of the earth's magnetic force, i=magnctic moment of magnet, W weight of magnet and ap- 

 pendages (compensation bar, mirror, stirrup, and part of suspension), 2 and -b the distances of 

 the threads /iborc and below, and /=length of suspension, then 



now let R and ~ vary bv SH and Ss and the ratio, " , or the variation of the horizontal force cx- 



a. 



pressed in parts of the force, is given by the relation 



/7 /f=cotz.te 



n 



"Suppose the scale division to be 1 millimeter and the distance of the scale and mirror=r 



i 



millimeter, then ~= l( , Now putting for a~ its equivalent a (n ii e ), where a = value of one division 



of scale in terms of radius and = the difference of any two scale readings, and making 



^ Jl 

 l;=<t cot :. the ratio, " . becomes /; ( ). A second method for determining the scale value is 



// 

 ii- 

 as follows: Let ic= , or let it be equal to any other convenient fraction of \Y, and add ic to the 

 1 0* ' 



AII important addition to tlir I'.rooko. instruments, as insuring tlm staliilily ortixity of tin- ilirrctinn ul'tln- ;'u 

 point of tin* scale; the idea \v;i taken from the later Ad ie magnet o^r.-ipli. Tlu> rireniar windows of I he I hive magnet - 

 oiiK-teis \vere D|' French ]i]ate-j;las.H. Ky trial on ]''<>l>ru;iry II. H J I. I lincl thai the transmitted rays Cor the rrlrfmr 

 scale-ends Hiifleied lint slight refraction by turning tlio f;l:iH.s in it< own plane; the deviation changed from II to 5 

 di \ i-iitc.H i n ma \ nin>. 



