more slowly and are of smaller dimensions than those of higher latitudes, but that they are, as a rule, of greater intensity. For the effect of rotation being less in low latitudes than in high, the distance from the centre of the disturbance to its edge will be much less.

It is not intended to imply that cyclones always move in the manner described above. Islands, continents, and mountain ranges, in conjunction with seasonal influences and those areas of high pressure called "anticyclones," these will greatly influence their movements. In fact, it is only the largest and the most rapidly moving ones that ever recurve eastwards and reach the higher latitudes. Moreover, many may have their origin in comparatively high latitudes, and many are simply offshoots from the intra-tropical disturbances.

The laws which govern the translation of the cyclones of lower latitudes have been very extensively dealt with in many works treating of that subject. It need only be briefly stated here that since the diurnal oscillations of the barometer in lower latitudes have been perfectly well ascertained, any even trifling anomaly in the behaviour of the mercurial column, taken in conjunction with any unusual behaviour of the wind over those oceans which are, at the season of the year when cyclones are common, ever visited by these devastating circulations, requires to be carefully observed by the navigator. A great access to his knowledge, and a great addition to his security, will be made

when the study of the cloud-forms in the neighbourhood of the intra-tropical cyclone has received the attention which it deserves.

147. The indraught of the lower winds into the central area of the cyclone of lower latitudes, while depending on the steepness of gradient and amount of frictional resistance presented by the earth's surface, is usually greater than the indraught into the cyclone of higher latitudes. Whether this indraught is greater in the front than in the rear of the disturbance, on the contrary, is a question which has not up to the present · time been satisfactorily answered. The author inclines. to the belief, which, however, requires confirmation, that the indraught is somewhat greater, or, in other words, the "angle of deviation" is somewhat less in the front than in the rear even of the cyclone of low latitudes. Above these cyclones the isobaric surfaces, greatly depressed near the floor of the atmosphere, are evidently considerably raised; and they are apparently more equally raised round all sides of the cyclone than is the case with similar systems in high latitudes. They seem, however, to be most raised in the front segments, since a brisk upper current usually flows off the disturbance in its front, an upper current which is of very great importance, in the relation to the cloud-forms which it carries, in enabling us to forecast the probable advance of the disturbance, more especially as the latter itself moves with a slow velocity of propagation.

148. With a view to illustrating the motions and characteristics of the cyclones of high and of low

latitudes, I have selected two examples, one of which (Chart III.) occurred in July 1873 in the North Atlantic, and the available data for which were carefully worked out in Synchronous Weather Charts of North Atlantic and Adjacent Continents.

Chart IV. represents another cyclone which traversed the Arabian Sea in June 1885, and is derived from Daily Weather Charts to illustrate the Tracks of Two Cyclones in the Arabian Sea. This sea is much less frequently visited by these disturbances than the Bay of Bengal, and very much less so than the China Seas.

149. It has been already stated that some of the cyclones of higher latitudes are poleward extensions of those of lower. A great majority, however, of the cyclones which traverse the North Atlantic and Pacific, and a very great majority indeed of those which traverse the higher southern latitudes, are first developed on the poleward side of the tropical highpressure belts. Their primary stage resembles closely that of the intra-tropical cyclones; that is to say, they are first formed over an area whose barometric gradients are slight, and where there is a humid state of the lower atmospheric strata. As a general rule, to which, however, there are exceptions, a young cyclone grows into existence in the rear of an old one, and it is partly for this reason that the cyclones of higher latitudes often follow one another for a considerable period along approximately similar tracks, and in rather regular sequence. When we consider the high velocity with which many of these systems