Two

I am indebted to Dr. Eschenhagen for copies of the six curves. of these (those of the declination and horizontal intensity on April 27) are

Wilhelmshaven (Dr. C. Börgen), 53° 31' 52" N., 8° 8' 48" E. The traces on the magnetic curves consist of a slight broadening of the curves. The times read off are those of the beginning of the disturbance in each case.

Parc St. Maur (M. Renou and M. Moureaux), 48° 48' 34" N., 2° 29' 38" E. No trace of any disturbance exists on the magnetic curves on April 20. On April 27 two pulsations are perceptible on the declination curve, the first very feeble at 7h. 54m., the second more marked at 7h. 59m. P.M. The curves of the two components of magnetic force are apparently undisturbed. Two bars of copper with bifilar suspension, orientated Ñ.S. and

E.W., show not the least sign of any disturbance.

"The movement,'

M. Renou remarks, 'is therefore magnetic and not mechanical.'
FIG. 2-Potsdam: Horizontal Intensity, April 27, 1894.

Utrecht (M. M. Snellen), 52° 5' 9" N., 5° 7' 55" E. April 27, the magnetic diagrams show unmistakable traces. For copies of them I am FIG. 3.-Utrecht: Declination, April 27, 1894.

7 56 10 99
'There is a very

Kew (Mr. C. Chree), 51° 38′ 6" N., 0° 18′ 47′′ W. small but unmistakable movement in the horizontal force curve, and a simultaneous extremely slight suggestion of a movement in the declination curve. Careful measurements give 8h. Om. P.M. as the mean Greenwich time of the middle of the movement on the horizontal force curve, and 7h. 59m. P.M. as that for the declination curve. There is not the faintest trace of movement in the vertical force curve.'

Birmingham, 52° 28′ N., 1° 54′ W. The pulsations were first seen on April 27 at 7h. 59m. P.M. Between Sh. 1m. and 8h. 3m. 20s. the image passed the cross wire twenty times, giving an average duration of 14 seconds for each oscillation. Between 8h. 8m. and 8h. 10m. 2s. the same number of oscillations was completed, the average duration of each being 12.2 seconds. The amplitude was determined by adjusting the image of the disc of light so that at one limit of its movement its edge coincided with the crosswire of the telescope. At 7h. 59m. the range was equal to three-quarters of the diameter of the disc. The whole diameter, it was afterwards found, is equivalent to 0.98 inch of the scale, so that the trace of the disturbance on a photographic recording apparatus in the same position as the scale would have been 18 mm. in breadth. As the angular value of the scaledivisions had not been ascertained since the beginning of August 1893 a new determination was made on the evenings of May 16-18. The mean of twenty-four pairs of tilts of 2" is 6.66±08 inches of the scale. Thus, at 7h. 59m. the range was 0"-22. After this I believe it slightly increased until 8h. 2m. or 8h. 3m. At 8h. 5m. it was 0-16. It then rapidly and almost continually diminished, being 0'11 at 8h. 53m., 0"-08 at 8h. 7m., 005 at 8h. 8m., and 003 at 8h. 12m. The movement then became so small that it could only be estimated. It was about 001 at 8h. 14m., 0005 at 8h. 16m. At 8h. 17m. there was a single oscillation of 0"-015. From 8h. 18m. to 8h. 19m. the image was steady, but at the latter time the range suddenly increased to 0"-03, but diminished after a few oscillations, until at 8h. 28m. the image was steady again. After this time no movement so great as 0.003 could with any certainty be detected.'

In addition to the above records it should be stated that the magnetic curves on April 27 have been examined at the following observatories with a negative result: Coimbra, Greenwich, Lisbon, Madrid, Nantes, Nice, and Stonyhurst.

The difference between the distances of Athens and Wilhelmshaven from Atalante is 1,910 km., and the difference between the recorded times at the same places is about 12m. 55s. on April 20, and about 10m. 9s. on April 27. Assuming that these times correspond to the same phase of the disturbance, we obtain 2:46 and 3.14 km. per second respectively for the average velocities on these days. These give:

2

Time at epicentrum on April 20=5h. 16m. 25s. P.M., G.M.T.
277h. 45m. 40s. P.M.

Using these values of the initial time, we have the following table :'Owing to the lag of the mirror through the oil these estimates are probably less than the actual amounts. When the frame of the pendulum is tilted suddenly through an angle of 2", the image at first moves quickly, but during the first 15 seconds not more than half its total deflection is accomplished.

* Since the duration of the disturbance at Athens on April 27 was only 12 seconds,