taken place in the apparatus. There was then obtained ẞ=2.976 and

tan 31

a=86.87 with stan

i

The mean value of the two determinations is a=82.36. Professor Edlund obtained a=130.99. The ratio of these values is 1.59.

The electromotive force of this combination was also determined from readings with intervals of 45 minutes.

As the result of these experiments I obtained ẞ=2·156 and

The ratio of the two values of a obtained according to the two different methods is 1.26.

The determination of the thermoelectric force in the way above given furnished the following result:

Experiment 4.-Pair of wires No. 1.

In all the following experiments we shall take the deflections, reduced to the temperature-difference 10° C. and the conductingpower 160, as relative measures of the thermoelectric forces. The magnitude of this force of the combination iron-copper is

The quotient obtained by the division of the value obtained by M. Edlund (146-18) by this is 1.70.

The values above obtained of the thermoelectric and electro

86.12

motive forces have the ratio

=1.05. If this number be

82.36

1.70

multiplied by the reduction-factor = 1.07, the number 1.12

1.59

is obtained, which is exactly equal to the numerical proportion of the two forces determined by Professor Edlund.

is greater than that corresponding to the combination iron-copper. Since it may be assumed that the molecular state of the iron-copper wire is pretty constant, this signifies that the rod of crystalline bismuth had undergone some molecular change since its examination by Professor Edlund. This alteration had induced a diminution of the electromotive force.

The thermoelectric experiments gave the following result:

Professor Edlund obtained for the thermoelectric force the value 835-10, the ratio of which to 460-06 equals 1.82. For the iron-copper combination this ratio was only 1.70. thermoelectric force of the copper-bismuth combination had therefore, like its electromotive force, diminished.

Alloy No. 1 and Copper.

The alloy consisted of 8 parts bismuth and 1 part tin. The experiments to determine the electromotive force gave :—

From these we get 8=7·178 and

a=240.56 from Exp. 11,

The electromotive force of the alloy appears therefore to diminish with time.

The thermoelectric experiments gave the following result :

From the first experiments it follows that :—

No. 1. Reduced deflection =252·98

Consequently the thermoelectric, like the electromotive force, proved somewhat smaller in the later experiments than in those first made.

Alloy No. 2 and Copper.

The elements of this alloy were 4 parts bismuth and 1 part tin. The electromotive force was ascertained by the following experiments :

The value of ẞ=4.815; with this we obtain

a=139.45 from Exp. 16.

Experiments were also made with intervals of 45 minutes, and furnished the following result:-8=4.991 and

The ratio of the two values of a is 1.29, which is, within the limits of errors of observation, equal to the ratio found with the iron-copper combination. Hence the method with shortened time between two successive readings gives a sufficiently exact result.

With the thermoelectric experiments the following. deflections were obtained :

Experiment 19.—No. 1.