ON CUllVKS OK nECLIiNATION JIAGNKTOGUAl'II.S, lUC. 201 



result constantly obtained when very small pressures were employed, and 

 wliicli may Lave arisen from the effect of tlio remanent occluded gas, there 

 also appears to be a sudden diminution of the capacity at about O'lO mil- 

 limetre pressure. To make sure that this was not an accidental result, 

 the air was pumped out again when the pressure was about 3 millimetres 

 until it was reduced again to about O'l millimetre, when the same dimi- 

 nution of capacity at about 0'19 was again observed. Now although 

 these numbers are merely now given as a preliminary indication of the 

 results obtained by the Committee, there is this interest about them, as 

 has been pointed out by Mr. G. F. Fitzgerald of Dublin, that the values 

 obtained for the capacity between about 002 and 02 millimetre's pressure 

 bear a general resemblance to those obtained for the Crookes' force. 



O .... 



One difficulty met with in the investigation consists m an apparent 

 change in the capacity of the condensers B and G (partly, no doubt, aris- 

 ing from changes of temperature) from day to day. A similar difficulty 

 was met -with in the previous investigation made in Japan, but it was 

 overcome by making alternate measurements of the capacity of the closed 

 condenser, first with air, then with vacuum, then with air, &c., &c. In 

 the present case this is, of course, impossible, since on account of the large 

 internal capacity of the condenser A, and the considerable quantity of 

 gas occluded in so large a mass of aluminium, it takes several days to 

 obtain a vacuum of 0001 of a millimetre even, although, at the suggestion 

 of Mr. Gimingham, induction sparks from a large induction coil (not 

 shown in the figure), are kept passing between the two sets of aluminium 

 cylinders at all times that a measurement of capacity is not being made. 

 Probably the best method of procedure is that followed on August 28, the 

 last day of the investigation, viz., first obtain slowly a very perfect vacuum, 

 no measurements of capacity being necessarily made, then admit into the 

 pump, drop by drop, mercury, occluding air, and make, during a couple of 

 hours or so, a complete series of measurements of capacity as the pres- 

 sure rises from, say, 0"001 of a millimetre up to ordinary atmospheric 

 pressure. Such a set of experiments being performed several times would 

 probably give a fair indication of tbe curve for capacity. As it is also 

 extremely desirable that the experiments should be made with statical 

 charges of electricity, the Committee have had constructed a somewhat 

 modified form of Thomson's quadrant electrometer, which they also pro- 

 pose employing for the measurement of the specific resistance of gas at 

 different pressures — the second half of their work, which they have not 

 yet commenced. 



Comparison of Curves of the Declioyxtion 3fagnetographs at Kew, 

 Stonyhurst, Goimbra, Lisbon, Vienna, and St. Petersburg. By 

 Professor W. Grylls Adams, F.R.S. 



[Plates VII., VIII., and IX.] 



[A communication ordered by the General Committee to be printed in extengo among 



the Keports.] 



During the month of March, 1879, there were several very considerable 

 magnetic disturbances, and therefore there were several favourable oppor- 

 tunities for comparing the effects of magnetic disturbances at different 



