2. For the case in which both the current of heat and the lines of force are at right angles to the chief crystallographic axis, positive values are only obtained for the highest temperature and very small field-strengths.

Case 1 a, §3. Although this does not differ from Case 1, it was thought worth while to take measurements with the crystal in this position. The dimensions of the apparatus did not admit of an extended series of observations. It was necessary to increase the distance between the magnetic poles to 3 cm., and the strongest field then obtainable was 4,400 C.G.S. units. Table III. gives the results, and side by side are the corresponding data for Case 1, obtained from the curves of fig. 5. The two sets of results, as was to be expected, are in fair agreement.

Case 3, § 3. The results are given in Table IV., and side by side the corresponding values for Case 2 as obtained from

the curves, fig. 5. The latter values are very different from the corresponding values of the former.

*

R. Defregger has already made some observations of the longitudinal phenomena with crystalline plates. The mean temperature is not given, but was probably about 50°, since the fall in the plate was obtained by application of steam and water at room-temperature. He describes the phenomenon as change of thermoelectric force by the magnetic field, and records the percentage change.

For purposes of comparison I have reduced the results for two field-strengths to this form. The following are the data :

Crystallographic Axis

Lines of Force.

Crystallographic Axis 1 Current of Heat.} Case 2, § 3.

In this case two specimens were used by Defregger. The dissymmetry with reversal of the field was noted by him, and the values given above are the means for the two directions of the field.

§6. The Transversal Effect.-This was calculated in accordance with the formula used by Yamaguchi *, viz. :—

where the positive X-axis is the direction of the current of heat, q the electromotive force between two points on a line at right angles to this direction and distant B apart, and m a quantity depending upon the field-strength and temperature, but in the first approximation independent of the fall of temperature. The direction of q as found by v. Ettingshausen and Nernst † can be determined by means of the following rule:-"Suppose one placed in the direction of the lines of force so that they enter at the feet and leave at the head, and to be looking at the same time along the direction of the flow of heat, then the electromotive force is directed from left to right.' In the following an electromotive force in this direction is reckoned positive, as both v. Everdingen and Yamaguchi have observed forces directed in the opposite direction. In the measurement of the transversal effect it is to be noted that copper electrodes were used. It is impossible to place these exactly upon an isothermal line (the fall of temperature is in some cases as much as 124° C. per cm.). Hence there exists a thermoelectric force between the points before the magnetic field is excited. On putting on the field there will be a rotation of the isothermal lines and a corresponding change of the thermoelectric force. The readings were taken quickly in order to avoid this error as much as possible. Yamaguchi used electrodes of electrolytic bismuth. This diminished the thermoelectric force considerably, but the error due to rotation of the isothermal lines was

unavoidable.

The results for both Cases 1 and 2 of § 3 are given in Tables V. and VI., and represented in the curves fig. 6. The full-line curve represents the results when the crystallographic axis is parallel to the current of heat, and the brokenline curve when the axis is at right angles to the current of heat. The abscisse are field-strengths, the ordinates corresponding values of m.

E. Yamaguchi, Ann. d. Phys. i. p. 214 (1900).

† A. von Ettingshausen & W. Nernst, Wied. Ann. xxix. p. 343 (1886).

d.

k.

d'.

d.

k.

d'.

d.

k.

d'.

d.

k.

d'.

(1) (2) &c. refer to the order of the observations.