470 
Proceedings of the Royal Society 
the other, than in the interior of continents in the same latitudes. 
Now, since it is to the interior of continents, viz., Asia, Africa, 
North and South America, and Australia, that these remarks on 
the discussing of heights are intended to apply, the limits' of error 
of single observations, or groups of observations, of the pressure of 
the atmosphere, are much less than one accustomed to observe 
barometric fluctuations in Great Britain might be led to suppose. 
Hence, if the mean monthly sea-level pressure of the part of the 
earth’s surface where the observation is made be kept in mind, 
the difference between this pressure and the observed pressure will 
be a tolerable approximation to the true difference of pressure due 
to the elevation of the place. 
But a still closer approximation may be reached. All examina- 
tion of weather on a large scale shows, in the most conclusive 
manner, that barometric fluctuations are always attended with 
changes of weather of a well-marked and determinate character. 
Hence, conversely, if travellers kept a careful record of the weather 
some time before and some time after they made their observations 
of the pressure of the atmosphere, some idea could be formed as to 
whether the observed pressure was above or below the mean pressure 
of the season at the place. 
Thus, suppose that for some time before and after the observa- 
tion the weather was fine and of a steady character throughout, 
the nights not much colder and the days not much hotter than 
usual, the winds light, or if moderate, continuing in one direction, 
and the state of the sky with respect to cloud much the same from 
day to day, it might be assumed that the pressure was the average 
of the season. Observations carefully made under these conditions 
are entitled to be ranked in the first class, as being the most trust- 
worthy that can be obtained. 
But if the nights have been for a day or two colder, the days 
hotter (in the sun), the air drier, and the winds lighter, and calm 
weather more prevalent than usual, then it is probable that the 
pressure at the time of observation was above the average of the 
season. 
Again, suppose, in the north temperate zone, the air to have 
become warmer and moister, the sky clouded, rain to have fallen, 
and the wind veered from E. or S.E., by S. and S.W. to N.W., or 
