TROPICAL AND EXTRA-TROPICAL CYCLONES.

It has been too much the custom in meteorological books to treat tropical cyclones apart from similar disturbances in extra-tropical or temperate regions. We have made numerous researches on the subject in India, the China Seas, Japan, and Mauritius, and found that, though the general character of all cyclones is the same, there are differences of detail which throw an immense amount of light on the cause of the great variety in the appearance of the sky in different parts of the same cyclone.

All cyclones agree in the great features of the wind rotating round the centre with a variable indraught, and of an upward and outward circulation of the higher

currents.

No more conclusive proof of this can be found than the fact that cyclones often pass out of the tropics, and then join or coalesce with others which have been formed without the tropics. Two similar eddies can easily unite, but two that rotated on different systems would infallibly break each other up.

The typical cyclone in all parts of the world is certainly oval, with the inner isobars usually closer to the rear than to the front; and the rain extends further before than behind the trough. But the tropical cyclone has a striking feature which is absent in our latitudes. There is a patch of blue sky over the calm centre, which is well known in hurricane countries as the "eye of the storm," or as a "bull's-eye." Then cirrus and halo appear all round a tropical cyclone, while they are never seen in

rear of a European storm; and though the way in which the rain seems to grow out of the air in front of a cyclone is the same everywhere, the sky and clouds in rear of a hurricane are much softer and dirtier than in temperate cyclones. There is not that sharp difference between the quality of clouds in front and rear which is so striking in higher latitudes. Still greater is the absence of any marked squall or change of weather during the passage of the trough in the tropics-that is, at the moment when the barometer begins to turn upwards. Some who study hurricanes have scarcely noticed any change then; and all are agreed that the trough-phenomena are very slight.

We have already shown, in our chapter on Prognostics, that a cyclone has, as it were, a double symmetry. One set of phenomena, such as wind, cloud, and rain, are grouped round the centre; while the second set, such as the different character of the heat and clouds in front and rear, and the line of squalls along the line where the barometer begins to rise, are related to the trough of the cyclone. If we call the first set the rotational, and the second set the translational, phenomena of a cyclone, we find that the former are all more marked in the tropical, and the latter in extra-tropical, cyclones. Then, if we examine the charts of cyclones, we see that, while tropical hurricanes are much smaller, and have much stronger winds than any others, they only move from two to ten miles an hour; while extra-tropical cyclones rotate much more slowly, but are propagated at a rate of from twenty to seventy miles an hour.

Thus we might readily suppose that what we call rotational phenomena are really due to the circulation,

and the translational phenomena to the forward motion, of a cyclone; and we are confirmed in this view by an examination of Japanese typhons. That semi-tropical country is traversed by cyclones of two different types at different seasons of the year, that move with different velocities, and they find that all the trough-phenomena are more marked in the quickly moving cyclones than in those whose progress is slower.

These researches also lead to another most important conclusion-that the character of cloud and weather depends on the position relative to the front of a cyclone, and not on the direction of the wind. Cyclones in Europe move towards the east, and the dirty sky comes with a south-east wind; while in the northern tropics hurricanes move towards the west, and the same sky comes with a north-west wind. People sometimes say that of course the rear of a cyclone must be clear, because of a cold, dry north-west wind; but when a cyclone moves west, even in Europe, that wind becomes close and dirty.

We shall defer our consideration of cyclones in the southern hemisphere till our chapter on Winds, because it is only the direction of the wind, and not the sequence of weather, which is altered in comparison with northern storm-systems.

ANTICYCLONES.

If we turn to the diagram (Fig. 16) of surface and upper winds in an anticyclone which we gave in our chapter on Clouds, we shall see at once that it presents

some analogies, as well as some very striking contrasts, to the cyclone figure. The anticyclone blows round and out below, round and in above, and therefore the conclusion is obvious that the air in the centre of an anticyclone must be descending. It must, then, necessarily be unusually dry, and this is just what observation shows it is. Then exactly the same argument holds as in a cyclone, that, as a whole, an anticyclone is a complex vortical system which possesses so much stability that great diurnal changes of temperature do not affect it as a whole.

It may be well to note the higher character of the explanation of weather which we can give now, as compared to what we said when treating of prognostics. Then we merely said that, as a matter of blind observation, the centre of a cyclone was rainy, and that of an anticyclone bright. Now we show that these two varieties of weather are the necessary product of different kinds of atmospheric eddies.

PRESSURE OVER CYCLONES AND ANTICYCLONES.

Simultaneous observations at the top and bottom of high mountains have demonstrated that the difference of pressure for a given height is always less in cyclones than in anticyclones; also that the fall of the barometer is always less pronounced on the summit than at the base of a mountain. For instance, if the difference of pressure between a high and low level station was four inches in a cyclone, it might be four and a quarter in an anticyclone; and if the barometer fell an inch at the sea-level,

the fall might only be about eight-tenths of an inch on the top of a mountain five thousand feet high.

The inference which is drawn from this is that, as we ascend, the gradients between a cyclone and its adjacent anticyclone must diminish, and it is by no means improbable that if we went up high enough we should find them inverted; that is to say, that the higher pressure would be over the cyclone.

In Fig. 20 we have drawn an ideal sketch of the

FIG. 20.-Probable vertical gradients over cyclone and anticyclone.

probable so-called vertical gradients over a cyclone and its adjacent anticyclone. The line at the bottom represents the level of the earth's surface. If the pressure is 31 ins. over the anticyclone, and only 27 ins. over the cyclone, the vertical isobar of 27 ins. must be as we have drawn it. But, as pressure decreases more rapidly over an anticyclone, it seems probable that at a certain level