120 
barometer was lower than it had been 48 hours previously, 
and 19 when it was higher. Forty-four when the barometer 
stood lower on the day of the explosion than the mean of 
the month in which the explosion occurred, and 35 when 
it stood higher. Sixty-six explosions occurred when the 
barometer was below the mean of the ten years, and 13 when 
it was above the mean. A further analysis of the table 
shows that taking the first pair of columns there were six 
explosions when the barometer, on the day of explosion, was 
less than a tenth of an inch lower ; and ten when less than a 
tenth of an inch higher than on the preceding day. When 
the difference was one-tenth, and less than two-tenths, 
twenty- two explosions took place with a falling, and 8 with 
a rising barometer. With a difference of from two to three- 
tenths 17 were with a falling, and two only with a rising 
barometer. Above three-tenths, 14 with a falling, and not 
one with a rising barometer. The law also holds good when 
applied to the other columns. 
The numbers in the third pair of columns would seem to 
indicate that the mean height of the barometer was generally 
low in those months when the greater number of explosions 
took place; and those in the fourth pair of columns also 
support that view. 
The second table shows the number of explosions which 
occurred in each month of the year respectively, and the 
mean differences found between the barometer readings on 
the days of explosions, and at other specified times. 
This table clearly shows that a generally low and falling 
state of the barometer exists when explosions take place in 
coal pits ; for there are in the whole series of numbers only 
two in which the positive sign indicates that the barometer 
was rising at the time of the explosion, and these are found 
in the column giving the comparison between the state of 
the barometer on the day of explosion, with the mean of the 
