— 231 — 
a decrease to an amount of 6 7, during the latter part. 
The whole range being thus 15 >. 
But some doubt prevails with regard to the reliability of 
the scale-value of the H-curve as registered at Karang-Sago, 
because the variation deduced from the H-curves for the 
period 9—17 May, so strongly exceeds the movement observed 
at Batavia on the same days. 
This. discrepancy, however, exists only for the morning 
hours, the amplitude between the maximum of H about noon 
and the afternoon-minimum showing. a fair agreement at both 
stations, just as the amplitudes of corresponding waves of 
short duration. 
Attention being den to this fact after the instrument 
was already dismounted, it is impossible to find out the cause 
of this disagreement, especially because the lightspot was often 
moved to some other part of the recording sheet by a very 
slight change in the angle of torsion of the suspension thread. 
Registering the H-curve on the days previous to the eclipse 
my object was not to obtain full curves for the 24 hours 
of the day, but only for the day hours. For the hours of the 
night the object of registration was to study the phenomenon 
of pulsation. 
Notwithstanding this uncertainty the scale-value of 1 mm. 
— 1.0 y used for the readings during eclipse-time must be 
reliable, as is shown by the following corresponding readings 
at Karang-Sago and Batavia: 
Karang-Sago, May 18". 
OP 49" a reading 128 mm., temp. 26.°70. 
db p » BER oos o LAE 
Difference 46 mm. 
Temp. eorr. —4 „ 
49 mm. = 42 y 
The Batavia-curve gives a range of 39 y and the small 
difference of 3 y may be ascribed to the difference in site 
of the two stations. 
