downward currents in the air. The upward currents are in areas over which a cyclonic circulation of the air exists, which in this Southern Hemisphere means a circulation with the hands of the clock, and screwing upwards. The upward motion of the air diminishes its pressure, and so the barometer falls. The downward currents are in anticyclonic areas, where in this Southern Hemisphere the circulation of the air is in the direction opposite to that of the hands of the clock, and screwing downwards, so increasing its pressure and showing a higher barometer.

The upward and downward motions of the air are performed in very complex swirls, and it is the algebraic sum of the pressures in all the superincumbent strata of air which is indicated by the barometer at any place and time.

A belt of cyclones appears to encircle the earth, having their centres ordinarily in about latitude 50° S. They average about fifteen hundred miles in diameter, and travel from west to east at an average speed of 250 knots in twenty-four hours. These cyclones are sometimes in parts of their courses shifted in latitude very considerably, and their rate of travel eastward is very much interfered with by anticyclones at times. Always their course and outline is very much interfered with by land, and especially by high land lying across their course as the Southern Alps of New Zealand.

The typical form usually taken up by the northern part of a cyclonic storm as it reaches the west coast of the South Island of New Zealand is shown on Plate XV. (a tracing from the isobars of the 11th May, 1900). The special effects of this diversion of the regular circular or elliptical form of the storm is to produce an eddy near Cape Farewell Spit, with northerly winds on the west coast and southerly winds on the east coast of the South Island, which continue southerly on the south coast of the Strait and bend round to northerly reaching Wellington from the north and bending round again to pass up the south-east coast of the North Island as south-west winds. Moreover, very often an anticyclone is developed at the same time in the great loop over the South Island. Thus it happens that what I had hoped would be an admirable position for the balanced wind-vane to give information about the up and down currents in passing storms, owing to its perfectly free open exposure, very often gives no indication whatever, owing to the eddy or calm between the beginning and the ending of the great bend or loop which encloses the South Island. Occasionally, however, when the centre of the storm is unusually far north, as on the 3rd October, 1899, it passes through Cook Strait, and immediately after the lowest point is passed the balanced windvane shows a change to a down current and the barometer

begins to rise, thus indicating that the up current is in the advancing part of the cyclone, and that the rising air is filled in behind from the following anticyclone. Further observations will be needed to make this quite certain, but I think there can be no doubt about the fact, although only this one instance has occurred during the year past in which the balanced wind-vane has been observed at Fare-well Spit.

In my paper of the 14th March, 1899, I gave what appeared to be the general system of circulation of the atmosphere in this Southern Hemisphere. The meteorological observations made by the late antarctic expeditions, by the Belgians and by Mr. Borchgrevinck, have shown that in the winter season a high barometer of over 30 in. is frequently observed as far south as 75° latitude. This indicates clearly that the very low barometer, hitherto always observed in the summer in those regions, is not a persistent fact resulting from some hidden cause, but that it is dependent on a belt of cyclones in about 75° S. latitude, as I have supposed, and that in the winter this cyclonic belt is sometimes displaced nearer to or farther from the pole (probably the former), and its place is occupied by anticyclones temporarily. We often see a similar result in the winter season here. The northern belt of anticyclones is displaced southward, and very high barometric readings are prevalent in the South Island. This has been the case in a very marked degree during this last winter. One result of this displacement this year has been, that the peculiar clouds which I have called “fish clouds,” and which are usually the precursors of a storm at Wellington, have very rarely been observed this year; south winds have also been unusually prevalent.

The constant formation of high anticyclones over the region of the Southern Alps of New Zealand in winter is a fact which has not yet been explained. A similar occurrence is observed over the high mountain belt of Norway, and it is probable that the low temperature of these great mountain masses is a cause of the descending currents of air over them.

To summarise, then, the results of my investigations into the motions of the atmosphere of this Southern Hemisphere, we find—

(1.) That the storms which reach us from the south-west are complete cyclones which circle round from north to east and from south to west while travelling round the earth from west to east, and having their centres usually about latitude 50° S.

(2.) That the old navigators’ “roaring forties” or the “brave west winds” are the northerly parts of these cyclones,

and that south of 50° of 55° S. latitude in the southern half of their circuits easterly winds are encountered.

(3.) That further south still anticyclones usually exist, and beyond them, in about latitude 75° S., a belt of cyclones cause the very low barometer usually observed there in the summer months.

(4.) That the occasional displacements of the belts of anticyclones and cyclones farther north or south than their normal position which we observe here appears to be also the case with the alternate belts of cyclones and anticyclones which exist between us and the pole.

(5.) That in cyclones there is an upward movement of the air, and in anticyclones there is a downward movement, each being accomplished in many complicated swirls, so that sometimes the inclination of the motion is, in gusts, as much as 40° from the horizontal, although the general upwards or downwards motion is very much less.

(6.) That it is the total effect of the upward or downward motions in the strata of the atmosphere above any place, at any time, which varies the atmospheric pressure shown by the barometer at that place.

(7.) That, while the barometer shows accurately the decrease of atmospheric pressure caused by the ascent of a mountain in calm weather, this mode of ascertaining heights is much interfered with in high winds owing to their deflection upwards by the slopes of the mountain.

(8.) That cyclones and anticyclones are complementary one to the other, and that apparently the rising air in the cyclone is replaced chiefly from the falling air in the following anticyclone.

(9.) That when cyclones collide, if moving in nearly the same direction they coalesce, if moving in nearly opposite directions they repel one another.

(10.) That occasionally small cyclonic circulations are originated in the neutral zones between two anticyclones; these move eastward round the more eastward of the two anticyclones.

(11.) That the cyclones which sometimes reach New Zealand from the north, with easterly winds at first, are tropical cyclones, originating in tropical regions northward of the belt of anticyclones, apparently caused by a concentration of heat in some place. Their forward motion is irregular, but it generally is in a southerly direction, and inclining to the east as they pass between two anticyclones. It is these tropical cyclones which sometimes meet one of the regular antarctic cyclones, and either blend with it or they repel one another, according to the direction in which the tropical cyclone is progressing at the time of meeting.