3. The strength or force alone of wind from any quarter (moisture and direction being unchanged) is preceded, or foretold, by a fall or rise, according as the strength will be greater or less, ranging, in an extreme case, to more than two inches. Hence, supposing the three causes to act together--in extreme cases--the mercury might range from about 31 (30.9) inches to near 27 inches, which has happened _occasionally_. Generally, however, as the three act much less strongly, and are less in accord--ordinary varieties of weather (the wind varying as usual--with more or less cloudiness, or rain) occur much more frequently than extreme changes. Another general rule requires attention; which is, that the wind usually veers, shifts, or goes round, _with the sun_, (right-handed in northern places, left-handed in the southern parts of the world,) and that, when it does not do so, or backs, more wind or bad weather may be expected instead of improvement. In a barometer the mercury begins to rise occasionally before the conclusion of gale, sometimes even at its commencement, as the equilibrium of the atmosphere begins to be restored. Although the mercury falls lowest before high winds, it frequently sinks considerably before heavy rain only. The barometer falls, but _not always_, on the approach of thunder and lightning, or when the atmosphere is highly charged with electricity.[16] Before and during the earlier part of serene and settled weather, the mercury commonly stands high, and is stationary.[17] Instances of fine weather, with a low glass, occur exceptionally, but they are always preludes to a duration of wind or rain, _if not both_. After very warm and calm weather, rain or a storm is likely to occur; or at any time when the atmosphere has been _heated_ much above the usual temperature of the season. Allowance should invariably be made for the previous state of the instrument during some days as well as hours, because its indications may be affected by remote causes, or by changes close at hand. Some of these changes may occur at a greater or less distance, influencing neighbouring regions, but not visible to each observer whose barometer, nevertheless, feels their effect. There may be heavy rains or violent winds beyond the horizon, out of view of an observer, by which his instruments may be affected considerably, though no particular change of weather occurs in his immediate locality. It may be repeated, that the longer a change of wind or weather is foretold by the barometer before it takes place, the longer the presaged weather will last; and, conversely, the shorter the warning, the less time whatever causes the warning; whether wind or a fall of rain, hail, or snow, will continue.