668 Sir J. J. Thomson on 



oE five solutions of widely different- concentrations, and these 

 be plotted on squared paper 100 cm. by 75 cm. ; on passing 

 the ruler through the points, interpolation values will be 

 obtained which, when tested b} r experiment, will be found to 

 be only 1 or 2 parts in 10 5 000 out; 



In a case such as this it would seem that theoretically the 

 Accuracy obtainable is limited chiefly by the unavoidable 

 errors in the paper (errors of ruling and of "stretch") and 

 iri the quality of the ruler* It is founds however, tuat a 

 practical limit is imposed by the weight of the apparatus ; 

 for, with the 1*5 metre ruler, the ease of adjustment is 

 already interfered with by the friction between the steel 

 edge and the paper. This difficulty may be overcome by 

 attaching vertical rods to the T's underneath the points V 

 and W (fig* 1), so that by the aid of springs the ruler is held 

 a millimetre above the surface of the paper-, but can be 

 depressed into contact by pressure. The lower end of the 

 rods terminates in a steel ball which slides on the polished 

 surface of three others above it, but rolls on the paper ; this 

 device is found to be a very efficient unlubiicated castor. 



LXV. Multiply-charged Atoms. 

 By Sir J. J. Thomson, DM., F.R.S.* 



[Plate XV.] 



IN the photographs of the positive rays (see, for example, 

 those given in the Phil. Mag. Aug. 1912) the mercury 

 line is remarkable for the exceptionally sinall displacement 

 of the head of its parabola. Even when the electric and 

 magnetic fields are strong enough to produce deflexions of 

 several millimetres in the heads of the parabolas correspond- 

 ing to the other elements, the head of the mercury parabola 

 is so little deflected that at first sight it seems to coincide 

 with the origin. When, however, the electric field used to 

 deflect the particles is made very large, in our experiments 

 from 5000 to 10,000 volts per centimetre, the head of this 

 parabola is distinctly displaced, and on measuring the electro- 

 static displacement it is found to be 1/8 of the normal 

 displacement of the heads of the parabolas Corresponding to 

 the other elements. 



The displacement due to the electric field is inversely pro- 

 portional to the kinetic energy of the particle displaced, so 

 that the atoms which produce the head of the mercury 



* Communicated by the Author. Eead before the Mathematical 

 Congress at Cambridge, August 26, 1912. 



