444 



Special Reports 



difference; it must be instrumental From 1911 to 1917 neither the piers nor the 



surroundings have been changed in any way. We have, moreover, computed the provisional diurnal 

 means for all the days of absolute observations of 1917. The seven first months are extremely 

 even, that is, up to the time of the accident. In September we find again precisely the previous 

 value: afterwards H increases a little. By applying to the first months the reduction on I. M. S. 

 and to the following months a correction which accords with the instruments of Mr. Brown, when 

 reduced to I. M. S., our last three months show a perceptible increase but not an unreasonable one." 



Thus it appears that at least some part of the change indicated by the present 

 comparisons is caused by change in constants arising from the accident. It is inter- 

 esting to note however also the gradual increase in the value of the correction on I. M. S. 

 as successively determined in 1907, 1911, and 1917. 



i All values by magnetometers No. 9 and No. 49 are referred to D; Da = F l'.O the weighted mean value from the 

 following determinations: September 1911, -l'.l (referred to new azimuth of mark at Da), weight 1; November 1917 by 

 magnetometer data as above, 0'.3, weight 1; and October to November 1917 by magnetometer No. 9 and magnetograph 

 scalings as above, 1'.3, weight 2. 



Table 12B Results of Horizontal-Intensity Comparisons at Hie Lukiapang Observatory, 1917. 



' All values are referred to Da\ > = /'' 2. S-. 



Half Bet; weight 0.5. 



