234 
MR. A. W. RtTCKER A^sD DR. T. E. THORPE OX A MAGNETIC 
line.s. A disturbance of the first class is however an obstacle to any inference as to 
the relation between the true and the terrestrial curves. It may distort them 
similarly, and so lead to false conclusions as to their undisturbed directions, or it may 
introduce such widespread and important irregularities that any conclusion deduced 
from the data afforded by the survey would be manifestly uncertain. 
It is to be noticed that the same cause may produce disturbances of different 
classes in the lines corresponding to different elements. Thus a broad band of 
magnetic rock at right angles to the magnetic meridian, the extremities of which lay 
outside the district under investigation, would yrrocluce little effect on the Declination, 
but might affect the Dip to a much larger extent. 
If an equation can be found to a family of curves which represents, as closely as 
possible, the general direction of the true isomagnetics of a particular kind throughout 
the whole country, the curves are the nearest approach to the terrestrial lines which 
can be deduced from the observations. The undulations of the district lines oir each 
side of these are evidence of regional disturbances, while the still more sinuous lines 
obtained by joining the points calculated as corresponding to any particidar value of 
the element from the values in two neighbouring stations are influenced by the local 
disturbances also. 
If the district under survey is very small, the assumption that the terrestrial lines 
are straight is very approximately true, but the nature of the curve indicated by a 
straight line on a map depends on the projection on which the map is drawn. 
The English observers have generally determined the position of the station by its 
distance north or south of a particular line of latitude, and east or west of a 
particular meridian, both distances being expressed in geographical miles. That this 
system is open to objection is evident by considering its application to a simple ideal case 
in wdiich the geographical meridians are themselves supposed to be the isogonal lines. 
It would evidently be better under such conditions to take the latitude and longitude 
as coordinates, and inasmuch as in the neighbourhood of the United Kingdom there 
is an approximation to such an arrangement, it is probable that the latitude and 
longitude are at least as convenient as any other data for the determination of the 
position of the stations. By this plan also, we are saved the trouble of converting 
the longitude east or west of Greenwich into miles east or west of the prime meridian. 
We have therefore taken the latitude and longitude as our coordinates. 
In the next place it is, we think, important to find the general equations to the 
isomagnetic lines, but it would be difficult to determine their form with sufficient 
accuracy by a graphical process. We therefore decided to trace them by a pre¬ 
liminary process of calculation which was carried out as follows. 
The country was divided into nine overlapping districts bounded by lines of latitude 
and longitude. If the stations within any district were not uniformly distributed, they 
were weighted, so that the weighted number of stations per unit of area should 
be everywhere about the same. In speaking of the mean value of any quantity in a 
