ant iel enn 
on the given date, while at 5.18 the descent had already 
set in. the movement of the needle slightly defleeting from 
the normal course later on in the day too. Kor this reason 
it is for the present impossible to state what the undisturbed 
course would have been. 
Two ways of building up a hypothesis on this are open: 
either two points lying on two rather undisturbed parts may 
be conneeted with each other — (2 casu those for 4.30 and 
4.0 are the most expedient tor this purpose — or the mean 
curve may be shoved parallel to itself in such a wav that 
the point for 5.18 shall coïncide with the corresponding 
one for 5.18 of the course for Mav I8®. If, however. 
there is an eclipse-disturbance. we have to fear that the 
one at 5.18, ie. half an hour before the time of totality. 
will have made itself felt alreadv. For Butaria the difficulty 
is of less importance; for here the curve between 4.30 and 
5.30 is pretty parallel to the mean one (9—17 May), aud 
it is obvious that the thing to do is to make the points for 
4.53 (12.0 Bat. t…) coïncide. It will then appear that the 
points for 8.0 again fall close to each other. which all the 
more supports our hypothesis. 
With respect to the Karang Sugo vector-curve of May 18'*, 
the first thing that strikes us. is the movement towards the 
west after 5.18 and its castward change of direction at 5.45. 
in consequence of which a loop is formed. 
The secondary turning-points fur 
horizontal intensity at 5h 45" 
declination at oh 45e 
(totality at Ker. Sago 5h 3844") 
are obvious on the magnetogram. 
After this the movement towards the east continues, in 
consequence of which a much larger loop is formed, in which 
the direction of motion is reversed, the maximum eastern 
elongation being reached at 6.55. 
From 8.0 to 8.30 the movement is pretty normal, but 
after this another deflection towards the west sets in. 
