158 



NATURE 



[Dec. 19, 1889 



draughts of air of different temperatures and with a lamp 

 and scale such as are used with a galvanometer, the 

 effect of the attraction can easily be shown to a few, or, 

 with a lime-hght, to an audience. To obtain this result 

 with apparatus of the ordinary construction and usual 

 size is next to impossible, on account chiefly of the great 

 disturbing effect of air currents set up by difference of 

 temperature in the case. The extreme portability of the 

 new instrument is a further advantage, as is evident when 

 the enormous weight and size of the attracting] masses 

 in the ordinary apparatus are considered. 



Fig. 



However, this result is only one of the objects of the 

 present inquiry. The other object which I had in view was 

 to find whether the small apparatus, besides being more 

 sensitive than that hitherto employed, would also be more 

 free from disturbances and so give more consistent results. 

 With this object I have placed the apparatus in a long 

 narrow vault under the private road between the South 

 Kensington Museum and the Science Schools. This is 

 not a good place for experiments of this kind, for when a 

 cab passes overhead the trembling is so great that loose 

 things visibly move ; however, it is the only place at my 



disposal that is in any degree suitable. A large drain- 

 pipe filled with gravel and cement and covered by a slab 

 of stone forms a fairly good table. The scale is made by 

 etching millimetre divisions on a strip of clear plate glass 

 80 centimetres long. This is secured at the other end of 

 the vault at a distance of io53"8 centimetres from the 

 mirror of the instrument. A telescope 132 centimetres 

 long with an object-glass 5* 08 centimetres in diameter, 

 rests on V's clamped to the wall, with its object-glass 

 360 centimetres from the mirror. Thus any disturbance 

 that the observer might produce if nearer is avoided, and 

 at the same time the field of view comprises 100 divisions. 

 While the observer is sitting at the telescope he can, by 

 pulling a string, move an albo-carbon light, mounted on 

 a carriage, so as to illuminate any part of the scale that 

 may happen to be in the field of the telescope. The 

 white and steady flame forms a brilliant background on 

 which the divisions appear in black. The accuracy of 

 the mirror is such that the millimetre divisions are clearly 

 defined, and the position of the cross-wire (a quartz fibre) 

 can be read accurately to one-tenth of a division. This 

 corresponds to a movement of the mirror of almost 

 exactly one second of arc. 



The mode of observation is as follows : When all is 

 quiet with the large masses in one extreme position, the 

 position of rest is observed and a mark placed on the 

 scale. The masses are moved to one side for a time and 

 then replaced, which sets up an oscillation. The reading 

 of every elongation and the time of every transit of the 

 mark are observed until the amplitude is reduced to 3 or 

 4 centimetres. The masses are then moved to the other 

 extreme position and the elongations and transits observed 

 again, and this is repeated as often as necessary. 



On the evening of Saturday, May 18, six sets of readings 

 were taken, but during the observations there was an 

 almost continuous tramp of art students above, producing 

 a perceptible tremor, besides which two vehicles passed, 

 and coals were twice shovelled in the coal cellar, which 

 is separated from the vault in which the observations 

 were made by only a 4i-inch brick wall. The result of 

 all this was a nearly perpetual tremor, which produced a 

 rapid oscillation of the scale on the cross-wire, extending 

 over a little more than i millimetre. This increased the 

 difficulty of taking the readings, but to what extent it 

 introduced error I shall not be able to tell until I can 

 make observations in a proper place. 



In spite of these disturbances, the agreement between 

 the deflections deduced from the several sets of observa- 

 tions, and between the periods, is far greater than I had 

 hoped to obtain, even under the most favourable condi- 

 tions. In order to show how well the instrument behaved, 

 I have copied from my note-book the whole series ot 

 figures of one set, which sufficiently explain themselves. 



0-805 

 0-808 



0807 

 0-807 

 0-805 



0-806 



0-802 



0-808 



o-8o8 

 0-814 



0-8066 



36-18 

 36-20 

 36-21 

 36-20 

 36-22 

 36-21 

 36-22 

 36-24 

 3624 

 36-26 

 36-26 



H.2 



9 

 II 

 12 

 13 

 15 

 16 



17 

 19 

 20 

 21 



25-0 



45-5 



5-3 



25-8 



45-0 



60 



25 o 



46-0 



4-5 

 27-0 

 44-0 



W.1 



80-2 

 80-2 



80 -o 



79-9 

 8o-i 

 80-1 



79-S 

 80-5 

 79-8 

 80-5 



80 08- 



