any actual depression, still many points, which have been noted, are so interesting that we shall mention some of them more in detail.

TENDENCY TO FOLLOW CERTAIN TRACKS.

During the persistence of any type, two or three successive cyclones have a remarkable tendency to follow the same course. This, of course, is the natural product of the fact that the path of a cyclone is determined by the type of pressure in which it is formed. Sometimes. this path is entirely dictated by surrounding pressure; but at other times local configuration of the land exercises a most powerful directive influence.

For instance, in Great Britain, during the westerly type, when the depressions are so far south as to cross that island, the centres have a decided tendency to traverse either the line of the Caledonian Canal in Scotland, or the low-lying ground which separates the valleys of the Forth and Clyde. Both of these courses coincide with what we may call lines of least resistance, for these are the two easiest lines by which it is possible to cross the mountainous districts of Scotland. Another wellmarked tendency of cyclone-centres is to hug the seashore, rather than to strike inland. When a cyclone comes up the English Channel, it often skirts the south coast of England, and then moves more northward along the east coast, rather than pass directly to the north-east across the land. In like manner, large cyclones which come in from the Atlantic, when they meet the coast of Norway, often hug the coast for several days, instead of

going straight to the north-east. In the United States the great majority of cyclones traverse the line of the great lakes, and then either follow the valley of the St. Lawrence or strike across the New England States into the Atlantic.

Great chains of mountains also influence very powerfully the paths of cyclones.

In Europe, the chain of the Alps almost forms a natural boundary between the weather of the Mediterranean and that of the northern portion of the continent. As a rule, that great inland sea has a totally different atmospheric circulation from that which affects the rest of Europe.

This will be very obvious if we turn again to the large charts which we gave in our chapter on Weather-Types.

Sometimes we can trace a cyclone in the Mediterranean trying to cross the Alps, and being broken up in the attempt. We can readily understand that if a mountain chain, 12,000 feet high, sliced off the lower half of such a shallow and complex vortex as a cyclone, the whole system might very easily be destroyed. Exceptional cases, however, do occur in which large cyclones cross the great barrier of the Alps.

In India, too, the still loftier chain of the Himalayas imposes an even greater influence on the meteorology of that country, as a glance at the charts which we have already given of the monsoon districts will abundantly show.

STORMS CROSSING THE ATLANTIC.

But the cyclones whose motions have created by far the greatest interest in Europe are those which sometimes

come across the Atlantic. The public have been fascinated by the idea that a storm could be telegraphed from New York, and its arrival on the coasts of Europe foretold three or four days in advance. If cyclones only moved with tolerably uniform velocities and in tolerable uniform paths, and the intensity remained constant, then, indeed, it would often be possible to obtain timely warning from the United States or Canada. Although the diagrams which we have already given of Atlantic weather would sufficiently show the real character of Atlantic cyclones, still the nature of the paths of these depressions will be more clearly understood if we give the tracks of all the depressions which appeared in the Atlantic during a single month. This will do as a sample of any other month or season. In Fig. 91 we therefore give a chart of all the cyclones which could be traced for more than two days in the United States, the Atlantic, and Europe during the month of July, 1879. During that month there were seven well-defined cyclone-tracks within the above-mentioned area.

These paths are plotted on our chart, and the position of the centre of each cyclone on every day is clearly marked.

Now, the first glance will at once satisfy us as to the broad idea that cyclones usually move in a certain general direction.

The whole of the paths lie along a comparatively narrow belt of the ocean; but when we come to look into the details, we shall find that the smaller variations of motion effectually preclude the use of this knowledge in forecasting.

Of the seven cyclones, four-Nos. I., II., V., and VII. -were formed in mid-Atlantic, and then pursued a more or less irregular course towards Europe. Observe how the curious loop to the northwards, which the path of No. I. makes at the beginning of the month, is almost exactly reproduced at the end of that time by cyclone

FIG. 91.-Cyclones crossing the Atlantic.

No. III. Cyclones IV. and VI. were formed over the United States; both passed into the Atlantic, but neither reached the coasts of Europe.

Cyclone No. III. also had its origin in the American Union, though, unlike the two others, it not only survived its journey across the Atlantic, but, after traversing Europe,

passed into Siberia. Our chart follows its history for the ten days from July 9 to 28; but let us try to see how we should fare if we attempted to issue forecasts on the supposition that the depression would move either in a uniform direction or with a uniform velocity. From the 9th to the 11th the cyclone moved towards the northeast with a considerable velocity; the next two days it turned to the south-east with diminished speed, and left the shores of the United States with a south-easterly trajectory. The day of leaving the velocity increased; but by next morning the direction changed again to the north-east, and the velocity gradually diminished for the next seven days, by which time the depression had reached the coast of Ireland, after being eight days in transit from Nova Scotia. A crack steamer would have done the distance in five days. From that day, the 21st, the speed increased again, and the cyclone turned still more towards the north. Then, with gradually decreasing velocity, the path bent round to the south, and afterwards turned once more to the northwards, with increased speed, till the 28th of July, when we lose sight of the depression in the frozen marshes of Siberia.

This example will abundantly prove that we can form no estimate of the future path or velocity of a cyclonecentre by any observations on its earlier motion. In this case the direction and velocity of the depression when it left the American shore gave no clue either to its path across the ocean, or its meanderings after reaching the continent of Europe.

There is another point which we must remember in the discussion of this question-we track cyclones, but