207 
pit, and the record sent to some party who might compare 
the observations with others taken either in other pits, or by 
parties keeping a regular register, at a considerable distance. 
Such comparison of registers would produce an efficient check 
on the man who had to record the barometric variations. At 
least four observations should be taken during the twenty- 
four hours, and whenever any fresh set of men went down 
to work, so that they might receive timely caution of danger. 
The method of recording the variation should be as 
simple as possible. A slate or piece of board should be 
furnished, ruled with columns and spaces for each time of 
observation, and on which the height of the barometer might 
be marked with chalk, and, for simplicity, it might be advan- 
tageous to use Roman numerals, and to make O represent 
28 inches, X 29 inches, XX 30 inches, and a simple stroke 
I for each tenth up to half an inch, then the V, and for 
one-twentieth a short stroke, thus i ; and on a third line 
mark the number of tenths by single strokes by which the 
barometer, at any observation, is lower than the preceding 
one; thus making the amount of fall appear evident with- 
out necessity for much calculation. Thus, if at six o'clock 
in the morning the reading were 30.25, at noon 30.40, at 
six o'clock in the evening 29.90, and at midnight 29.40, 
the table would be as follows : — 
6 
XX.IIi 
Noon. 
XX.IIII 
6 
X. Villi 
inn 
12 
X.IIII 
mil 
