AND INSTRUMENTS AT MUNICH. 503 
the wooden framework of the building. A constant current of 
air between the two walls can be maintained, and regulated so 
as to prevent damp ; and in fact none of the oppressive sensation 
of damp often experienced in under-ground buildings is sensible. 
At 1 2 3 4 are wooden pillars, destined to support the magnetic 
instruments; these pillars were saturated with linseed oil, and 
are as yet in perfectly good preservation. M is the insulated 
pillar which supports the theodolite. At the eastern extremity 
A, a perpendicular shaft has been sunk to 30 feet below the sur- 
| face of the ground. A ladder leads to the bottom, and long 
thermometers are placed in the east side of the shaft in such 
) manner that the bulbs are imbedded 2 feet deep in the earth, 
the projecting part of the tube being turned up so that the height 
of mercury is read off on the vertical portion of the tube. These 
thermometers are placed at every 4 feet from the top to the 
bottom of the shaft, with a view to the investigation of the 
propagation of the sun’s heat through the soil. The subterranean 
passage E F communicates by steps with the library of the 
astronomical observatory. At the western extremity there is a 
quadrangular opening 6 at the same height as the theodolite at 
M, which is prolonged under ground in a straight line until it 
issues on the south-western declivity of the rising ground on 
_ which the observatory is built; through this opening the theo- 
dolite views a distant mark on the church of St. Anne, in a 
suburb of Munich, the azimuth of which has been accurately 
determined. The passage E F is covered with earth, but the 
observatory itself is not; the roofs of the side buildings rise about 
2 feet above the ground ; that of the middle is still higher, and 
has glazed openings above M, by which dayhehi is admitted 
into the observatory. 
The exact instruments introduced by M. Gauss, so far sur- 
passing the accuracy before attained, appeared to open com- 
pletely anew the question of the manner in which the investiga- 
tions of terrestrial magnetism ought to be conducted. There 
are two classes of phenomena for the study of which observa- 
tions of measurement are employed, viz. such as being called 
forth by one or by a small number of causes only, may be fol- 
lowed and explained through all their variations ;—and such as 
have besides their principal cause, an infinite number of sub- 
ordinate accidental causes, so that it is only in their mean values 
that their subjection to laws can be traced out, whilst an expla- 
