998 REPORT—1896. 
Each level is 2°50 metres long, and consists of two vessels A, B (fig. 17) 
30 cm. in diameter and 25 cm. high, communicating with one another by 
means of a tube C, 15 cm. in diameter. The level is filled with water, 
and, to prevent evaporation, floats, D, E, consisting of zine dishes 28 cm. 
in diameter with a rim 3 cm. high, are placed in the vessels at each 
end, and these again are surrounded with a stratum of oil. In the centre 
of the float D there rests a weight F of 100 grammes, connected by a wire 
with the end of the short arm of the amplifying lever G, the fulcrum of 
the lever being fixed to a plate H resting on top of the vessel A. The 
arms of the levers are 3 mm. and 15 em. in length, so that the movements 
of the float are magnified fifty times. The longer arms of the levers were 
at first furnished with pens filled with ink, but for these were afterwards 
substituted points writing on smoked paper, which give much clearer 
diagrams. The paper is wrapped round a cylinder K, rotating on a 
vertical axis once in 53 minutes, and drives the paper under the pen at 
the rate of 55 mm. a minute. The levers of the two levels are arranged 
with their pens on the same vertical line, and about 6 em. from one another. 
In order that the records may not be superposed after a complete 
revolution is made, a cylinder L, 4 cm. in diameter, is lowered from a 
drum, driven by another clock, into.the vessel B. As it becomes immersed 
in the water the registering float D is slowly and gradually raised, and the 
pen in consequence traces a continuous spiral on the paper. As the 
Fie. 17; ® 
te 
cylinder rotates once in 53 minutes the diagram for each component 
consists of twenty-seven lines, the distance between consecutive lines 
being about a millimetre. To determine the time of any displacement of 
the float, a trace is impressed when the paper is put on and taken off, as 
well as at some intermediate time about equidistant from the ends of the 
24 hours. Or, if desired, an automatic hourly trace could be made by 
electric connection with a chronometer. 
The lines of the spiral are parallel and equidistant. Except when the 
instrument requires sensitising, the registra- 
Fie. 18. tion proceeds without jumps, showing that it 
is sensible’ to very small changes of level. 
Et Professor Grablovitz informs me that he 
has not been able to determine the smallest 
tilt which the levels can detect, but a displace- 
ment of the writing-point of half a milli- 
metre, corresponding to a tilt of the ground of 
2” can generally be read with certainty. 
The levels do not seem to be affected by 
the tremors of passing carts, &c., but they are 
sensible to certain seismic movements. They 
will not register slow movements taking place 
in a horizontal direction, for in such cases the 
water receives no displacement relatively to 
the tubes. Nor do they seem capable of 
recording the long-period pulsations of very distant earthquakes. For 
instance, on June 15 of the present year the horizontal pendulums at 
2st a4™ 
= 
= 
= 
a) 
a 
~ 
