236 
between them in the indications of the barometer and ther- 
mometer, yet the quantity of rain on one of the days nearly 
double what it is on the other.—This must be due either to an 
increased amount of vapour brought in suddenly by south-westerly 
currents at the elevation at which condensation takes place, or to 
a lowering of the temperature at that same elevation by the influx 
of a northerly current, neither of which circumstances would 
necessarily, or at least immediately, affect any instruments near 
the ground—not even the vane, though relatively so much higher 
than the other instruments, for reasons already stated. 
Sometimes again, on two consecutive days, each with some rain, 
when the humidity on both days is at saturation point, the 
temperatures of the air and dewpoint being therefore coincident, 
and when these temperatures are also absolutely the same on both 
days, with little difference too in the barometer, a much larger 
fall of rain is found on one of the days than on the other, owing 
to a change of wind below ;—the cause and its effects being here 
the same as in the last-mentioned instance, only the former 
at once manifested by the vane. 
Many other instances of a similar kind might be brought 
forward from the daily registers. It is noteworthy also that 
several of the monthly summaries show the like discrepancies 
between the results of the instruments, and the general character 
of the weather as regards wet and dry. 
Thus January, 1869, had a rainfall exceeding five inches, and 
more than double that of January, 1872, yet the mean height of 
the barometer was very nearly the same in both months ; in fact 
that of January, 1869, which had most rain was the higher of the 
two. But the mean temperatures of the air and dewpoint were 
each of them three and a half degrees higher in 1869 than in 
1870, the humidity too being 90 (saturation=100) in the wet 
January, and only 86 in the drier January. These differences 
quite explain the circumstances above stated. 
If now we look to the month of December, 1872, and compare 
q 
} 
5 
