60 Scientific Proceedings, Royal Dublin Society. 



approximation, and if we go no farther there will remain but one 

 arbitrary member in the entire of the resulting series of systema- 

 tic physical units. 



If we also introduce the numerical value found above for C 

 which depends on E^ and is less accurately known, we obtain the 

 following actual values for these units of Nature. 



Lj = j^^^^jj of a metre (13). 



Tj = o y T V "^ ^ second (14). 



Mj = ^yT^ of a gramme (15). 



or, in other words — 



The natural unit of length approaches in value to the thirty 

 seventh-metre, [i.e., the metre divided by 10*''). 



The natural unit of time approaches in value to one-third of 

 the forty fifth-second, {i.e., one-third of the second of time 

 divided by 10^^); and 



The natural unit of mass approaches to the seventh-gramme, 

 (i.e., the gramme divided by 10'^). 



This appears the best attempt we can yet make to determine 

 these remarkable units. In the series to which they belong all 

 the electrostatic units will be identical with the corresponding 

 electromagnetic units ; all the forces of Nature that are known to 

 obey the law of the inverse square, whether they arise from gravi- 

 tation, electricity, or magnetism will be expressed without 

 coefficients ; and the chemical bond which seems to be the unit 

 of concrete Nature is brought into its proper relation to physics. 



Postscript. — Many persons find it difficult to conceive of Gj as a unit. 

 Gi may be avoided and Mj substituted for it, if Mi be defined as a mass 

 such that it attracts an equal mass at a distance with the same force 

 with which two imits of electricity as defined on p. 56 (i.e. each equal 

 to Ej), would, if placed at the same distance asunder, act on each other. 

 The three fundamental units of the Natural System will then be Yj, E^, 

 and Mi, from which all others are to be derived. This M^ is the same 

 as the Ml of equation (15). 



