1 



4 On the general distribution of Terrestrial 



constant, the known increase of the dip would account for a 

 rather larger diminution of the horizontal component ; and since 

 it is probable that the total intensity is likewise slightly on the 

 increase, the result obtained from our scanty data may be con- 

 sidered siiflEiciently well established to be used. 



Our knowledge of the secular changes on the Western Coast 

 and in the territories is so deficient that no satisfactory reduction 

 can be applied to the observations. The changes, however, are 

 known to be small, and the observations do not differ greatly in 

 date. Their mean date is about 1852, which may be considered 

 as the period to which the western part of our map corresponds 



more nearlv than to 1850. 



Construction of Maps, — In the construction of the lines on the 

 maps, both the graphic and analytical methods have been used. 



Observations within limited spaces were united into groups, 

 by taking the arithmetical means; a number of such groups 

 were combined by conditional equations of the second degreCj 

 amounting to an interpolation by second differences. 



The several systems of groups were so arranged as to overlap, 

 and the slight disagreement in the joining was adjusted by an 

 interpolation^ partly graphic and partly arithmetical. 



When the latitudes and longitudes of stations have appeared 

 unsuitable co-ordinates of position, owing to the stations being 

 distributed in an oblique direction to the meridian^ they were 

 carefully projected on a map, and referred by measurement to an 

 assumed axis of co-ordinates in any convenient linear measure ; 

 the lines deduced from the conditional equations so formed, being 

 projected according to the same system, the latitudes and lon- 

 gitudes of points in them could be read off and tabulated, the arti- 

 ficial system serving only as a convenient means of interpolation. 



On the accompanying maps the lines of equal declination, dip^ 

 and horizontal intensity have been drawn only as far as they 

 were warranted by observations. For places within the range 

 of the lines approximate values may readily be obtained by 

 graphical interpolation. 



Table IV gives the differences between the observed values at 

 the Coast Survey stations, reduced to 1850, and the correspond- 

 ing values on the maps. It will be seen that there are a few 

 large residuals, sometimes exceeding 2° in declination, 1° in dip, 

 and 2V of the horizontal force, which belong to isolated stations, 

 (see Mt. Pleasant, No. 5, and Patuccawa, No. 15,) or very limited 

 localities (see stations near Cape Ann, Nos. 20 to 24) ; in these 

 cases local attraction is too apparent to allow the observations to 

 be used in the construction of the map. 



We find, further, that in certain more extended localities the 

 residuals in declination, amounting in the average to about 20' 

 or 25\ have one sign, indicating a more general deviation from 



a regular system. Thus, the observed declinations near New 



