1813.] Scientific Intelligence, 219 



with the air it is gradually changed into a black pov*'der (the 

 protoxide). This must happen, to a certain amount, In the 

 marine transit, from the continual running of the mercury, 

 unless the air be excluded, which seems impracticable. Two 

 methods of remedying this defect (which would interfere with 

 the rate at which the mercury flows) present themselves. The 

 first method would be to fill the whole apparatus with some gas 

 destitute of oxygen, and as nearly as possible of the specific 

 gravity of air. Azote is the gas that would answer best ; but I 

 doubt whether it would be practicable to confine azote in any 

 such apparatus for a great length of time, to the complete 

 exclusion of air, considering the great changes in expansion to 

 which it would be liable. The other method is to have aa 

 apparatus for filtering the mercury through leather. By 

 repeating this filtration once a week or so, and taking care to 

 keep the apparatus and mercury always as dry as possible, 1 have 

 little doubt that it might be preserved nearly in a constant state 

 of purity. How far time may be accurately measured by tHe 

 flowing of mercury, remains to be decided by experiment. It 

 is well known that water is far from flowing equally ; and 

 accordingly the clepsydras, or water clocks, of the ancients, were 

 of little value. Mercury certainly promises fairer; but 1 much 

 doubt whether it be possible to regulate its rate of flowing so as 

 to make it absolutely equable. 



VIII. Variation of the Compass, 



Though the fact, that the compass is constantly changing the 

 direction towards which it points, has been known since the year 

 1645, when it was discovered by Gellibrand, or Mair, in Eng- 

 land, yet we still know nothing about the laws according to 

 which it varies. To give our readers some idea of this variation, 

 we shall set down a few examples of the v<^;ation in London, 

 during different years : 0, or no variation, means that the needle 

 points due north. The figures indicate the number of degrees 

 that it deviates from north : — 



1580 , . ,11° 15' 0'' East . . . .Burrows. 



1622 6 Gunter. 



1634 4 3 30 Gellibrand. 



1657 0 Bond. 



1672 2 30 West Halley. 



1683.. .. .... 4 30 Halley. 



1692 6 Halley. . 



1722 14 20 ...... ..Graham. 



1747 17 40 Graham. 



1811 ,,24 14 2 Lee. 



