of Mass, Length and Time. 
157 
The thermoelectric phenomena are given by the points where 
the lines of the thermoelectric diagram cut the axis of tempera- 
tures and their slope. The first quantities are not very small. 
The slope has its maximum value for nickel, ‘24 microvolts per 
degree per degree or, in the new units, 2‘6 x 10 -2 . 
8. The diagram represents these numbers on a logarithmic 
scale, numbers greater than unity being represented by marks 
above the line, while those less than unity have their reciprocals 
represented by marks below the line. As anticipated, the numbers 
1 10 io 2 10 3 104 105 v io 6 10 ^ IQ 3 10 9 
I Vi'i'I 1) i'll! "I 1 1 1 h-H 1 — H h 
1 jv 
Fig. 1. 
crowd together in the neighbourhood of unity, but the range is 
much greater than one would expect. This is not to be ascribed 
to the complexity of the phenomena causing large coefficients to 
enter into the equations expressing the quantities in terms of the 
fundamental properties of the ether, for the velocity of light, 
which can hardly be a complicated function of these properties, 
has the large value 42 x 10 5 . If this velocity be a fourth funda- 
mental property, we should expect a crowding of the points about 
this value, which we do not find. This, however, depends on the 
kind of unit used to measure electromagnetic quantities in. If 
we had measured them in electromagnetic units we should have 
had several cases of approximation to some power of v. We may 
possibly explain the large values of some of the quantities by 
supposing that they contain f v. There is no indication of any 
fifth fundamental quality of the ether. 
9. We may now take those quantities that lie between 100 
and and find what units will reduce them to small values. 
I find in this way the units 
l = 7’60 x 10 -8 cms., m — 2*34 x 10 -23 gms., t = 7 - 50 x IO -13 sec. 
These are very near the old units, and if we express the quantities 
anew in terms of these units none of the conclusions reached 
above will be altered. We may note however that the values for 
the mean mass of the atom and the charge on the electron are 
now T08 and 2'4 respectively, values which are practically unity. 
As both these quantities are of fundamental importance, it is 
highly confirmatory of the present theory that their new values 
should be near to unity. 
Another quantity of fundamental importance is the constant 
of gravitation, G'6 x 10 -8 or 2 - 0 x 10 -23 in the units of § 9. As v 
