112 REPORT— 1881. 



tunifcy was afforded for testing the disturber. At G.16 P.M. of that day 

 we began the readings. The changes from ' up ' to ' down ' were made as 

 quickly as we could, and in a quarter of an hour we secured five readings 

 when the guide weight was 'up,' and four when it was 'down.' 



When a curve was drawn, with the time as abscissa, and the readings 

 as ordinates, through the 'up's,' and similarly through the 'down's,' the 

 curves presented similar features. This seems to show that movement of 

 the disturber does not cause irregularities or changes, except such as it 

 is designed to produce. 



The displacement of the guide weight was through 5 cm. on each 

 occasion. 



The four changes from ' up ' to ' down ' showed that an inch of scale 

 corresponded with 0"-0897, with a mean error of 0"'0021 ; the four from 

 'down' to 'up' gave 0"'0909 to the inch, with a mean error of 0"-0042. 

 Thus the systematic error on the previous occasions was probably only 

 apparent. 



Including all the eight changes together, we find that the value of an 

 inch is 0"-0903 with a mean error of 0"-00r!0. 



A change in the scale reading amounting to a tenth of an inch is 

 visible without any doubt, and even less is probably visible. Now it will 

 give an idea of the delicacy of the instrument when we say that a tenth 

 of an inch of our scale corresponds to a change of horizon ' through an 

 angle equal to that subtended by an inch at 384 miles. 



II. On the luorJc of previozis observers. 



In the following section we propose to give an account of the various 

 experiments which have been made in order to detect small variations of 

 horizon, as far as they are known to us ; but it is probable that other 

 papers of a similar kind may have escaped our notice. 



In a report of this kind it is useful to have references collected 

 together, and therefore, besides giving an account of the papers which 

 we have consulted, we shall requote the references contained in these 

 papers. 



In Poggendorf's ' Annalen ' for 1873 there are papers by Professor 

 P. ZoUner, which had been previously read before the Royal Saxon 

 Society, and which are entitled ' Ueber eine neue Methode zur Messung 

 anziehender und abstossender Krafte,' vol. 150, p. 131, ' Beschreibung und 

 Anwendung des Horizontalpendels,' vol. 150, p. 134. A part of the 

 second of these papers is translated, and the figure is reproduced in the 

 supplementary number of the 'Philosophical Magazine ' for 1872, p. 491, 

 in a paper ' On the Origin of the Earth's Magnetism.' 



The horizontal pendulum was independently invented by Professor 

 Zollner, and, notwithstanding assertions to the contrary, was probably for 

 the first time actually realised by him ; it appears, however, that it had 

 been twice invented before. The history of the instrument contains a 

 curious piece of scientific fraud, of which we shall give an account below. 



The instrument underwent some modifications under the hands of 

 Professor Zollner, and the two forms are described in the above papers. 



• We use the expression ' change of horizon ' to denote relative movement of 

 the earth, at the place of observation, and the plumb-line. Such changes may arise 

 either from alteration in the shape of the earth, or from displacement of the plumb- 

 line ; our experiments do not determine which of these two really takes place. 



