usually commences with the appearance of fine streaky cirrus clouds, which gradually change to cirrostratus, and then the uniform coating of cloud is complete. This cirrus represents the equatorial current, seen from beneath, which has already set in above, and marks its progress by the streaks of cloud. The water, on its condensation from the state of vapour, assumes the solid form at once; so that these high clouds are not composed of bubble steam*, but of minute spiculæ of ice, and they give rise to the halos of the sun and moon, which are caused by refraction of the light, and to the so-called rings, mock-suns, and mock-moons.

If these appearances accompany a falling barometer, it is a sure sign that wet weather is coming on. The reason that the long streaks of the cirrus appear to us as arcs of circles which diverge from one point of the horizon and reunite at the opposite point, is that they are projected on the apparently curved surface of the sky. This apparent curvature of the cirrus differs from the lateral feathery off-shoots of the same clouds, which show that the direction of the upper wind is not quite constant. This latter form of cirrus, consequently, is a less certain sign of rain than the long arched clouds are. There is another form of cirrus, which does not always indicate rain, as the air gets warmed during the day and ascends; if the temperature be high this ascending current sometimes reaches as high as the cirrus above, and then the latter

* Nebelbläschen is the word translated 'bubble steam' (vapeur vesiculaire, Fr.). It indicates visible steam, as distinguished from vapour, which is invisible. (Trans.)

+ The hollow winds begin to blow,

The clouds look black, the glass is low;

Last night the sun went pale to bed,

The moon in halos hid her head:

'T will surely rain.

breaks up into small cumuli, which are known in Germany under the names of Schäfchen, Lämmer-Gewölk in South Germany, brebis in France, fleecy clouds in England, and were called by the Romans vellera lana. Howard calls them 'cirro-cumuli.' In the south of Europe they are said to be a sign of rain. In Northern Germany this is not the case, according to my observations. When the atmosphere is warm and dry, the outlines of distant objects become indistinct and hazy, owing to the dust which is suspended in the air, and the sun appears reddish. If easterly and northerly winds have lasted for a long time in summer, with very dry weather, and a moist wind sets in, its aqueous vapour condenses itself at once on the dust which is in the air, which thus becomes heavy and sinks to the ground. Under these circumstances, the air becomes very clear, and in mountainous countries the mountains appear quite close, and the waterfalls are heard more distinctly. This is considered an infallible sign of rain.

The rain comes, as a general rule, from the west side; so that a clear sunset is a proof that there is no rain coming from that quarter for some time. Hence, this is considered to be a sign of fine weather.

In the evening, when the air ceases to ascend, the clouds sink, and are dissolved in the warm strata below. From this nothing can be augured for the following day. There is an old French proverb

Temps, qui se fait beau la nuit,

Dure peu quand le jour luit.*

If the atmosphere is very damp, evaporation cannot go on, and this feeling produces in us the sensation which we designate by sultriness, drückende Luft. The direct

* Weather which clears up at night, will not last when the day breaks.

action of the sun is then more felt, and we say scorching.

the sun is

If the south wind sets in suddenly in the upper strata in winter, the rain falls at once at that level, and small transparent grains of ice, i. e. rain frozen while falling, reach the ground. We say then that Glatteis (glazed frost) is falling, as the rain which soon sets in freezes on the ground and glazes it. We may then expect a SW. storm with a great fall of the barometer.

In winter, rain, with a west wind and a rising barometer, turns to snow; snow, with an east wind and a falling barometer, to rain.

In spring, if the wind shifts through W. to N., we may expect the weather to clear up suddenly and night frosts to set in, even though the thermometer, at a little height above the ground, may not fall below the freezing point.

Heavy thunderstorms, which come up with an E. wind while the barometer is falling, do not cool the air. We say that it is still sultry and there will be another thunderstorm. The air does not grow cooler till a thunderstorm comes up from the W. and the barometer begins to rise.

If several thunderstorms come on in succession from the W., each successive storm is usually more northerly than that which has preceded it.

In the case of thunderstorms from the W., the under current is usually more northerly than the upper one; consequently the true thunderclouds (cirro-strati) drive more or less at right angles to the lines of the cirrus above them.

The greater the difference of temperature between the two currents which displace each other, and accordingly the greater contrast there is between their directions, the more likely they are to produce a thundercloud. The winds stop each other's path and produce a calm before

X

the storm comes up. The cold wind then breaks in suddenly, and it is a mistake to say that the thunderstorm has made the wind change.

The winter thunderstorms in Norway are westerly storms, in which the under current is shifting quickly towards N. The barometer rises and cold follows. They are preceded by a thaw, mild weather, heavy rain, and southerly winds.

Our winter thunderstorms in Germany, which are rare, exhibit the same character. There is, however, another type of these storms. This is exhibited when the equatorial current sets in with great violence. In such a case, the thunder and lightning are often so tremendous that we say the sky is bursting open (der Himmel öffnet sich).. They are followed by a complete spring wind.

The time at which thunderstorms are most common is regulated by the commencement of the rainy season. They are most common in the height of summer in the torrid zone; in mid-winter in the district of the sub-tropical rains at the outer edge of that zone; in spring and autumn in the south of Europe; and in the middle of summer in that part which is north of the Alps, with the exception of Norway. They are, on the whole, rare in the frigid zone, but yet do occur there up to high latitudes. Lastly, they occur in volcanic districts, as secondary results of the rapidly ascending current above a volcano in eruption, and at times at which they are never observed unless under these circumstances.

If bad weather continues for a long time, the vane oscillates between SW. and W., and the barometer fluctuates slightly. This is the true equatorial current.

Thunderstorms in spring lie at a low level, and do not last long; they are usually followed by a return of cold weather. They are at times accompanied by sleet or

snow, and they frequently do considerable damage by means of lightning. The lower wool-pack clouds drive with a WNW. wind, the upper cirrus with SSW.

If the barometer rises very quickly, this indicates, not that the southern and northern currents are interfering with each other laterally, but that they have met and mutually stopped each other's way. A severe storm is sure to follow; and if the barometer falls as quickly as it has risen, it shows that the southern current has prevailed, and that the danger is therefore close at hand. In this case, the lettering of a barometer which bears very dry' for this level is totally wrong.

6

If in winter a cold and a warm wind meet each other, and the southerly current has not sufficient force to overcome the resistance of the northerly current which opposes it, the barometer rises to a great height at the line of contact, and a thick fog appears there. This fog often disappears suddenly and reappears again, according as the southerly current gives way a little, and the place of observation comes off the line of contact into the true northerly current, and vice versâ. If severe cold follow such a fog, it shows that the northerly current has finally prevailed.

If the barometer at any place oscillates violently, and the air remains at rest there, the disturbance must lie in a lateral direction. At times in winter the southerly current prevails over a large area, and the barometer is low, the air delightfully mild. Under these circumstances there is a very severe winter, with a high barometer somewhere in the neighbourhood. It is possible that this cold air may force its way, as a storm, into the warm and rarefied air in its neighbourhood, and cause the barometer to rise rapidly.

On the Atlantic Ocean, if the wind veers against the