I append graphical representations of the rarefactions which are observed throughout the area where that phenomenon is most remarkable.

If we seek to determine the area in which the rarefaction reaches a maximum value in the northern hemisphere, we find it to exceed 0.35 in. over an area whose southern limit passes from Barnaoul to the shores of the Sea of Aral, then turns to the eastward so as to enclose the whole plateau of the Gobi, and even to embrace Shanghae and Pekin on the east coast of Asia, while Benares and Calcutta are all but included within it. The area in which the rarefaction is between 0-18 and 0.35 in. includes the eastern side of the Ural range lying to the south of Catharinenburg, the Kirghis Steppe, the Caspian Sea, Persia, Afghanistan, Southern Arabia, the catchment basin of the Indus, the valley of the Ganges, the northern portion of Burmah, and the south of China. The

August

boundary turns northwards, between Canton and the Philippine Islands, and passes through Japan to the coast of Mantchouria. The extreme limit of the area in the Pacific Ocean i.e., the point where the rarefaction ceases to be observable, is as yet undetermined. The western limit passes from the neighbourhood of St. Petersburg to the south, so that the west coast of the Black Sea, Asia Minor, Syria, and Egypt, are included in it. The point where this boundary meets the coast of Africa may be determined with considerable accuracy, inasmuch as Palermo and Tripoli lie outside of it, as may be seen from the following table : —

The stations in Algeria exhibit clearly the deflection of the barometrical curve, while those in the Azores and Canary Islands do not do so. It is, therefore, probable that the desert of Sahara is included within the area of rarefaction, as Abyssinia has been proved to belong to it.

The distribution of atmospherical pressure does not agree directly with that of temperature. In the month of

July the isothermal lines form a closed curve in Abyssinia and Arabia, enclosing an area in which the temperature reaches a maximum value. From this point the temperature decreases materially towards the district of maximum rarefaction in northern Asia. As regards the western boundary, the coincidence with the is-abnormals is much closer, but even this does not hold good on the east coast of Asia. If we remember that all the great deserts are included within the area, we are referred so directly to the action of the aqueous vapour that we are justified in expecting to find the grounds of this phenomenon therein.

In spring, according as the sun's meridian altitude increases, the temperature in Siberia rises very rapidly, so that the difference of temperature between northern Asia and Hindostan is materially decreased. It is still, however, sufficient to preserve the current of air as NE. Monsoon from higher to lower latitudes. After this period the barometrical difference is considerably diminished.

In Barnaoul, the tension of aqueous vapour does not exceed 0.18 in. from January to May, while in Calcutta it rises 0.35 in. On the other hand the barometer falls 0.35 in. in Calcutta during the same period, and in Barnaoul nearly 0.44 in. By the month of June the total fall in Barnaoul reaches 0·66 in., and in Calcutta only 0.49 in. We see from this that the point of minimum atmospherical pressure lies to the northward of that of maximum temperature, so that the current of air from the southward is attracted not only to this latter point, but beyond it. Another consequence of the rapid change of position of the point of minimum pressure is that the SW. Monsoon frequently sets in suddenly. Under other circumstances we should only expect the NE. Monsoon to set in suddenly, as it is a colder current than the other.

The long continuance of the SW. Monsoon is explained

by the fact, that this current, in its passage over the southern slopes of the table-land of Central Asia, parts with a great portion of the moisture which it contains in the copious precipitation of rain which there takes place, and consequently is insufficient to fill the vacuum which exists over that plateau. The monsoon is in nowise a phenomenon produced by the attraction of air from a colder towards a warmer point, analogous to the daily land and sea-breezes before described, as is still maintained by modern writers who follow in the steps of the old textbooks. Its cause is to be sought for in the temperate rather than in the torrid zone.

The direction of the monsoons is thus accounted for. The SE. Trade-wind owes its motion from the east, to the fact of its passage from points which possess a slow to those which possess a more rapid velocity of rotation. As soon as this current crosses the equator, conditions, which are exactly opposite to the above, occur, and the wind becomes SW. In precisely the same way the NE. Trade-wind, on crossing the equator, becomes the NW. Monsoon of the southern hemisphere.

2. THE WEST MONSOONS OF THE LINE.

This is a peculiar modification of the Trade-winds on the Atlantic Ocean, which is known on the coast of Guinea. Owing to the position of Upper Guinea, the SE. Trade-wind is attracted so far to the northward that rainbringing winds from the SW. and WSW. are felt over the whole district up to the Cape Verde Islands. Between this deflected trade-wind and the ordinary trade, there is interposed a region of calms, which is marked on the maps of Halley, Dampier, and Mushenbroek, as the 'Rain'

F