blowing from the same point several days in succession. It is only by daily weather-telegrams from the north of Russia that warning of such gales can be sent to seamen about to navigate the dangerous waters of the Pentland Firth. And since similar heavy gales prevail in North Britain, when a storm from the N.W. passes over the north of Norway and descends on the Baltic, the desirableness of weather-telegrams from the north of Norway, so as to complete a system of storm-warnings as applicable to Great Britain, will readily appear. Hence forecasts made of British weather without weather-telegrams from these localities must occasionally lead to inaccuracies, and give an empirical look to any system of forecasting, if these omissions in the weather-telegrams be not allowed for.

508. Storms do not always proceed in the same uniform direction from day to day, but occasionally change that direction. Though the change which occurs in the direction of the progressive movement is generally small, as shown by the tracks or blue lines on Plate VII., yet sometimes it is very considerable. Storms are sometimes observed to die out on reaching Russia, and when this happens the barometer at places to the eastward of the storm has been rising, instead of falling as usual in the front of storms, thus giving evidence of the descent of a polar current of dry heavy air over central Europe. In passing through Russia storms frequently have a more contracted area than they had in passing over western Europe-a remark of general application to storms in passing from a wet into a dry climate. Observations show that the two storms on the chart, Plate VII., died out, the one in Finland and the other in the Baltic.

509. The storms on the coasts of the Mediterranean follow a different course. Some proceed from the north to the south, influenced probably by the heated air rising from the Sahara; a number appear to come from the east, and pass in the direction of N.W. over, Greece and Italy towards the Alps, where their onward course is arrested; while others travel in an easterly direction,-some passing from the Bay

of Biscay up the valley of the Garonne, down upon the Gulf of Lyons, and thence to the east; and others passing south of Spain and then pursuing an easterly course.

510. Rate at which Storms travel.-Figures or dates will be observed on the blue lines, Plate VII., indicating the tracks followed by both storms. These show the position of the centre of the storm on the mornings of the successive days on which they appear on the charts. The dates are, beginning in the west, for the first storm, 29th, 30th October, 1st and 2d November; and for the second storm, 1st, 2d, 3d, and 4th November. These give a mean rate of about 18 miles an hour, which is nearly the average rate at which storms travel across Europe. Sometimes the rate is only 15 miles an hour; or even, as already stated, in certain situations and under certain circumstances, 0 miles, or stationary; whilst on rare occasions it is as great as 45 miles an hour. Since the distance from the south-west of Ireland to the east of Great Britain is about 450 miles, it follows that even after a storm has appeared on the west of Ireland, the eastern seaports may generally be warned of its approach twenty-four hours before the force of the gale begins to be felt, even though no warning be issued until the storm has actually made its appearance in Ireland.

511. The shortness of the time generally elapsing between the commencement of a storm on the west of Ireland and its arrival even on the eastern ports of Great Britain, points out the inutility of any system of storm-warnings which does not include the receipt, six or eight times a-day, of barometric and wind observations made every hour in the extreme west of Ireland. For thus only can the appearance of a storm likely to pass over the British Islands be detected in time, followed in its course by the telegraph, and timely warning of its approach sent to different ports; and thus only can security be had that the storm will not outstrip the "Signal" announcing its approach. It need scarcely be pointed out that this frequent transmission of observations from Ireland to London is not called for unless when a storm is expected.

Let the instruments there be in charge of one who has some knowledge of weather-changes,-of the bearings of the cirrus cloud, the direction of the wind, and the fluctuations of the barometer, in their relations to storms. Then, if everything appear settled, no additional telegraphic despatch is required; but, if the usual premonitions of a storm show themselves, he should at once telegraph the fact to London, and continue to do so from time to time as already indicated.

512. General Path of Storms over Europe.-This is a feature of storms which has not yet been investigated. To do so properly would require an examination of many storms at all times of the year, to show the influence of season, if any, in changing the line of the general route; and, from that examination, to lay down on maps lines similar to the blue lines on Plate VII. From those I have examined, I am inclined to think that the two blue lines show not only the two most general directions in which storms travel, but also the average position of the line of the routes-the track north of Great Britain being the more common. I have calculated the mean monthly range of the barometer at about eighty places in Europe, for the eight years beginning with 1859. The lines of equal annual disturbance lie in a direcfrom W.S.W. to E.N.E., of which the following indicate the position of a few :

In July these lines lie from south-west to north-east; but in winter the direction is nearly due east and west, the mean monthly barometric range in Orkney being 1.712 inches, and at St Petersburg, 1.797; at Dublin, 1.421, and at Königsberg, 1.474; and at Bayonne, .926, and at Trieste, .033.*

* As connected with this subject, see an interesting paper by Joseph Baxendell, Manchester, on the Periodic Disturbances of Atmospheric Pressure in Europe and Northern Asia. (Mem. Lit. and Phil. Soc. of Manchester, 1860-61).

513. Relations of the Temperature and the Dew-point to Storms. The temperature increases a few degrees at places towards which, and over which, the front part of the storm is advancing, and falls at those places over which the front part of the storm has already passed. In other words, the temperature rises as the barometer falls, and falls as the barometer rises. Generally the temperature in advance of the storm is above the average, and in the rear of the storm below it; but if it should chance to be considerably above the average in advance of the storm, it may still be above the average when the storm has passed, though lower than it was before. If the temperature begin soon and markedly to rise after the storm has passed, a second storm may be expected in a short time. In front of a storm the dew-point is high, in the rear it is low. This state of things is not what is merely due to the high temperature in the one case, and to the low temperature in the other; for, along with the high dewpoint there is great humidity, and along with the low dewpoint a small humidity. To express this in popular language -before the storm, the air is close and warm (for the season); but after the storm has passed, the air is cold and dry.

514. Relations of Rain and Cloud to Storms.-When the barometer has been falling for some time, clouds begin to overspread the sky, and rain to fall at intervals; and as the central depression approaches, the rain becomes more general, heavy, and continuous. After the centre of the storm has passed, or when the barometer has begun to rise, the rain becomes less heavy, falling rather in showers than continuously; the clouds break up; and fine weather, ushered in with cold breezes, ultimately prevails. The rainfall is generally proportioned to the suddenness and extent of the atmospheric depression at the place where it falls.

OBSERVATIONS OF THE WIND.

515. Direction of the Wind.-Let us first look at the winds in the storm which was passing over the British Islands on the 2d November 1863 (Plate VII.) On that occasion the least atmospheric pressure was in the central districts of England, from which it rose to some extent all round. The arrows, representing the winds exactly as observed at the different stations, show that the general direction was as follows:-W.S.W. and S.W. in the north, west, and centre of France; S. in north-east of France, in Norfolk, in Belgium, and in Holland; S.E. in north of England and south-west of Norway; E. and N.E. in Scotland and Iceland; N. in west of Scotland, and north and east of Ireland; N.W. and W.N.W. in west of Ireland and south-west of England. Thus the wind blew in all directions round the central patch of least pressure.

516. Again, if the direction of the wind be referred to the region where the pressure was least-viz., in the centre of England-it will be seen that it did not blow directly towards this region, but to a point somewhere to the right hand of it. Thus in France, south of the area of least pressure, the wind was not S. but W.S.W. and S.W.; in Belgium and to the south-east of it, the wind was not S.E. but S.; at Silloth, Solway Firth, to the N.N.W. of it, the wind was not N.N.W. but N.E.; and in the west of Ireland, to the west of it, the wind was not W. but N.W., &c. With regard to the isobarometric lines, which are coloured red on the chart, the wind did not blow round the lines in circles returning on themselves, nor did it blow across them at right angles, holding a straight course towards the region of least pressure, but it took a direction somewhere intermediate between these two directions.

517. This relation is what is known as Buys Ballot's "Law of the Winds." According to this distinguished meteorologist, if a line be drawn in the direction of the wind, and