462 E. W. Willson — Magnetic Field in 



be nearly uniform. The result would be that at points north 

 of the center the field would be weakened and at points south 

 strengthened, while in a line east and west through the center 

 of the pier the field would be undisturbed by its presence. 



The preceding table shows the variations of the Horizontal 

 Force actually observed along a horizontal line passing 7 inches 

 above the centers and edges of the piers 5, 6, 7, 8 in room 15, 

 fig. 4. 



Column 1 gives the distance of the point of observation 

 from the south wall. Columns 2, 3 and 4 the corresponding 

 difference in decimal parts of H, from its value assumed as 

 normal in those parts of the room farthest removed from the 

 influence of the piers. Columns 5 and 6 give the correspond- 

 ing values on parallel lines 14 inches east and west of the cen- 

 tral line. Columns 3 and 4 are introduced to furnish a means 

 of judging of the accuracy of the individual observations, 

 each of which rests upon a single reading of the instrument 

 in each of the positions, ~N. pole east and JST. pole west. 



Although the differences on different dates amount in the 

 maximum to *0007 H, owing to uncertainty in the diurnal vari- 

 ation of H, which has been assumed to be uniform during the 

 time o£ observation, it will be noticed that the difference be- 

 tween consecutive points (upon which the important features 

 of the disturbance depend) as observed on various dates, is the 

 same for 32 out of 39 points, within -0002 H, which is prac- 

 tical identity, since any observation may be in error in either 

 direction by '0001 H. In four cases the difference is *0003 H, 

 in two cases *0004 H and in one case *0005 H. This is remark- 

 able testimony to the accuracy of the instrument when it is 

 remembered that each determination required but little over 

 two minutes for its completion.* 



The results of column 1 are best exhibited to the eye by fig. 

 6, in which the abscissas represent distances from the south 

 wall in feet while the ordinates give the corresponding differ- 

 ences from the normal undisturbed value of H for the whole 

 room. The places of the piers are as indicated below. 



It appears that in passing from south to north the value of 

 H increases as we approach the pier 5 to a maximum almost 

 exactly over the edge of the brick-work and six inches inside 

 of the bluestone, then rapidly decreases to a minimum almost 

 exactly over the northern edge of the brick-work, from which 

 it increases, reaching the average value for the room at a point 

 about half way to the next pier. Beyond this it still increases 

 to a maximum over the edge of pier 6 and follows the same 



* For a detailed account of the method of observation see the paper by F. 

 Kohlrausch, Wied. Ann., vol. xix, p. 130. As my instrument was adjusted a 

 difference of one division in the reading of ni—n 2 corresponds to 00025 H. 



