THE SUPPORT OF THE HIMALAYAS. 101 



some seconds of arc, owing to the effect of the irregularity of 

 topography in the immediate vicinity of the station. The stations 

 furthest in the hills are situated on peaks, and are not so much 

 affected by this cause as those near the outer edge of the hills, where 

 the effect is considerable. The stations of Mussooree and Banog, 

 for instance, are situated on a ridge with a deep-cut valley on the 

 north, and would therefore show a southerly deflection as compared 

 with similarly situated stations on an imaginary representative of 

 an average Himalaya, and the difference in the observed deflections 

 at the two stations seems sufficiently accounted for by the local 

 topography, which makes the effect of the valley to the north 

 greater at Banog than at Mussooree. At Rajpur, which is situated 

 at the southern foot of this ridge, the effect of the valley to the 

 north is less, and here we have a northerly difference ; the mean of 

 the three gives a small southerly difference, or residual, if the 

 effect of the trough has been correctly estimated. 



The most conspicuous characteristic of the figures is the excess 

 of observed over calculated deflection in a northerly direction, 

 exhibited at all the stations in the interior of the hills, amounting 

 to from 8" to 12", and the smallness of the differences at the outer 

 stations, where the positive differences are as numerous as the nega- 

 tive. Though these characteristics are common to all the groups, 

 it will be well to examine each separately. 



In the western group we have first the two stations of Kidarkanta 

 and Lambatach, at a mean distance of a little under 40 miles from 

 the. main boundary and giving a difference, or residual, of northerly 

 deflection amounting to about 10"; next the two stations of Bahak 

 and Bajamara, at about 20 miles from the boundary, give a differ- 

 ence of 2* and 0" respectively, and thirdly the three stations of 

 Mussooree, Banog, and Kajpur, all within 3 miles of the main bound- 

 ary, give a mean difference of about 3" southerly. In all these 

 cases the differences depend in part on the effect of the trough, and 

 the dimensions adopted for this were those which have been deduced 

 as probable ones, namely, 10,000 feet depths under the Siwalik area 

 and 15,000 feet under the plains beyond. The adoption of these 

 figures was largely governed by the fact that tables had been calculated 

 for those depths, and the estimate is probably somewhat in excess of 

 reality ; if this excess amounted to as much as 25 per cent., probablv 

 an extreme value, the northerly deflections at the three outer stations 

 would be reduced by about 4" and the mean difference, or residual, 



[ 249 J 



