storm is a marked characteristic of areas of low pressure, but the indications of temperature have, for various. reasons, not yet been reduced to strict rules like those of the barometer.

We know, however, that a great contrast of temperature between adjacent stations or, so to speak, a great thermometric gradient, being an indication of serious atmospherical disturbance, is the precursor or concomitant of a serious storm, so that if we find a great difference between the reports of temperature, we are at once warned of approaching disturbance. A recent instance of this fact may be cited. On November 13, 1875, at 8 a.m., the thermometric report from Scilly was 57°, and from Wick 21°. The resulting difference is 36°. The gale of Sunday, November 14, with its accompanying high tide and consequent inundations on the south coast and in London, will be fresh in the memory of all.

Besides all the symptoms of the approach of a storm which have been enumerated, there are the innumerable local signs which enable the fisherman and shepherd, not to speak of the inferior animals, to judge of coming bad weather. Among all these, almost the most important are the character and motion of the upper clouds, showing the existence of wind aloft which will, in all probability, soon descend to our level. Another great sign is the clearness of the atmosphere, the unusual visibility of distant objects, which is well known to all as a sign of a coming gale.

It must always be remembered, that in order to issue perfectly correct storm warnings, we should require to know the size, shape, position, and motion, in direction and rate, of an advancing depression, and also whether it is becoming deeper or the contrary, and that there is not one of these conditions of which we have a really sufficient knowledge at present, while of most of them we can have no knowledge at all till the storm has burst upon us. The problem which is put before the Meteorological Office daily, is similar to one which astronomers would at once recognise as impossible of solution, and that is to determine the elements of a comet's path from a single observation taken, say, in a brief clear interval on a cloudy night. The first glimpse we get of a storm must suffice for us to issue our warnings.

It is therefore evident that for our own exposed western and northern coasts we can but rarely issue timely warnings, but fortunately these iron-bound shores are not frequented by an amount of coasting craft at all to be compared with that navigating the comparatively calmer waters of the two Channels and the North Sea.

The results of the warnings to our own coasts have been printed as Parliamentary papers for several years back. The following is the abstract of these results for the year 1874 (Parl. Paper, No. 210, 1875, being the last which has been published).

Return of the Result of the Comparison between the Warnings issued and the Weather experienced in 1874.

If these figures be compared with those for the year 1873, we see that the results in percentages are nearly identical.

Accordingly it appears that nearly half the signals of approaching storms (force upwards of 8, Beaufort scale, 'a fresh gale') were fully justified, and the same is the case for four out of five signals of approaching strong winds (force upwards of 6, Beaufort scale, a 'strong breeze').

It is therefore evident that the system of storm warnings renders it possible to give useful intelligence to the coasts, but it need hardly be said that the system for these islands is rather of the 'from hand to mouth' type, and that our warnings would be far more useful if they could be issued sooner, so as to render it possible for captains to get intimation of a coming gale before leaving port. This, however, cannot be hoped for in the present state, the infancy, of weather telegraphy, and a few instances will suffice to show what are the principal causes of failure of warnings, in addition to those arising from the imperfections of our arrangements which have been already described.

I shall commence with a case in which even a

may remember the Royal Adelaide' day at noon the

practised seaman's experience was at fault, so that we can hardly blame our telegraphic reporters, who are mostly landsmen, for their failure to notice from the appearance of the sky and sea, that a tremendous storm was imminent. Some of my readers gale of November 22, 1872, when the was lost on Chesil Bank. On that telegraphic reports showed an apparent improvement on the weather of the previous day, so that the signals then flying were lowered on the south coast. At night the storm came on, and of course the comments on the Office were not favourable. On this day one of our best sea-observers, the late Captain Thomas Donkin, of the 'Inverness,' (one of the three ships that rode out the Madras cyclone of May 2, 1872), was out in the gale, and was blown back by it, hove to, from the Lizard to the Casquets, off Alderney. I wrote to him as soon as I heard that he was safe at anchor at Portland, to ask him whether he, being at sea, had anticipated the storm from the look of the sky and sea, and his answer was: 'With respect to the weather on November 22, I may say that at noon I was standing in towards the land, between Falmouth and Plymouth, and a pilot cutter came alongside, and if I had had the least apprehension of such a gale as followed being near at hand, I should have taken a pilot and gone into Plymouth. The appearance of the weather at the time was fine, though the glass was falling, though not low at the time for SW. wind and unsettled weather.'