surface of the earth has perpetual calms. The calm-belts themselves are not constant, but move in the different seasons, and besides, the calms are more or less frequently disturbed.

In the trade-winds belts also, notwithstanding calms are very rare, the velocity of the wind is probably less than the average of the globe.

Probably the part of the earth where the winds have the greatest velocity, is found between 40° and 60° Lat. S., where very strong westerly winds are prevailing the whole year. The cause of this is the great difference in the pressure of the air at a small distance, or in other words the steep barometric gradient.

The great difference of the mean velocity of the winds blowing over a region, and of the progress of the air in a certain direction, should be borne in mind. Where the winds are weak, but always from one direction, as in the trade-wind region, the total rate of progress measured in miles will be considerable, frequently greater than in regions where strong winds blow from different directions. It is even possible that the winds may be so counterbalanced by one another, that there will be no resultant direction, so that the definite result, as far as progress of the air is concerned, would be the same as if absolute calms had prevailed all the time. So far as regions are considered, where the mean direction of the wind does not vary, or varies but slightly in the different seasons, the mean annual direction with rate of resultant, gives a tolerably fair idea of the character of winds in such regions.

It is quite different where regions with very great variations in the yearly direction of the wind are considered. Here the annual direction will give but a very imperfect idea of the character of the winds. This is the reason why, as far as possible, I have always placed at least two contrasting seasons, summer and winter, in giving the percentages of the winds and the mean directions in the small tables which follow, and serve to illustrate the winds of different regions of the world. This is also the reason for constructing the two maps, Pl. 5 and 6. The same attention has been given to this subject by Prof. Coffin in his extensive tables arranged in Zones, in Series B of this work, the number of observations being given generally for the four seasons, sometimes even for each month. How far the consideration of the annual result alone would mislead, the following table will show:

It is seen from the foregoing that in Western Europe, on the Atlantic Ocean and in a certain region of the United States, the mean direction of the wind in the year, summer and winter, is between S. W. and W. S. W., and the difference between the two seasons very small. If the rate of annual resultant is not greater, it is because at all seasons there are many winds coming from other directions than the prevailing one.

In Pekin and Hakodade the mean annual direction is nearly the same as at the above named places, but the ratio of resultant is small for another reason: the winds of summer and winter being nearly opposite to one another, the resulting annual movement is small, Yet at each of the seasons the winds are very steady. The angle between the mean direction of the wind in winter and summer is 142° at Pekin, and 133° at Hakodade, or more than of a circle, and only from 3° to 20° at the above cited places of Europe and America. Again, the mean annual direction of the wind and ratio of resultant, in Southern India and Ceylon, are very similar to those observed in Europe, but the mean direction of winter and summer nearly opposite to one another, with an extremely great ratio at both seasons, there are conditions as dissimilar as possible to those of Western Europe.

In the pages which follow, the results to be drawn from the observations on the winds are considered by geographical divisions.

SPECIAL DEDUCTIONS.

GREENLAND AND ARCTIC AMERICA.

THE information we have on the winds, as well as on the general climate of Arctic America and the adjacent islands, is more extensive than that on any other Arctic region, Northern Norway excepted. Our knowledge of these regions is mostly due to Arctic explorations. The Arctic Archipelago, north of the American Continent and west of Greenland, was explored almost continuously by British expeditions for more than thirty-five years (1818-1855), in search of a northwestern passage.

The results of these expeditions are of high value to science, especially as the inducements to explorations in this direction can scarcely ever return. The bays and straits between the islands are probably the most ice-bound in the world.

Smith's Sound and Northern Greenland have been explored by the American expeditions of Kane, Hayes, and Hall.

According to the most authentic Arctic authorities, Smith's Sound offers the best route to the Pole, the sea between Spitzbergen and Nova Zembla perhaps alone excepted. It is entirely frozen only a short time, and does not present serious obstacles to navigation in steamers. This gives us reason to expect further knowledge of those regions which were so successfully penetrated by American explorers, with very inadequate means at their disposal.

A German expedition wintered in Eastern Greenland, 75° N. Lat.

We know much less of Western Arctic America; few expeditions having wintered there west of 100°. Our knowledge of the interior of British America is also less than of the Arctic Archipelago, though it is much more easy of access. More information relative to this region is very desirable.

Our knowledge of the climate of Arctic regions generally having been mainly derived from observations made in the Arctic Archipelago of America and in Smith's Sound, it is necessary therefore to inquire into the geographical position of these regions. They are situated from nearly due north to W. N. W. of Iceland, where, as was stated above, exists the lowest pressure of the northern hemisphere, nearly the whole year round, but especially in winter. This must lead to the prevalence of northerly and westerly winds. Accordingly in the stations in Smith's Sound northeasterly winds were found dominant, owing to the influence of the strait, and also to the position, N. N. W. of Iceland. (See Map, Pl. 2.)

There are great discrepancies in the results obtained at the different stations,

but these are easily accounted for, if we remember that the period of observation was short, mostly one year only, and that the climate of the Arctic regions is very changeable; still there are some differences in the direction of the winds which can only be ascribed to their geographical position. Thus Northern Greenland has the greatest prevalence of the true polar winds, northeast, and this is due in no small degree to its proximity to Iceland, as well as to the open water of Smith's Sound near a very cold continental area.

The most northerly stations west of Smith's Sound, as Northumberland Sound and Port Refuge, have the least amount of northern winds. This is, no doubt, owing to their distance from Iceland, and, probably also, to a partly open sea to the northward of them. If there is really an open sea in this direction, the pressure there must be lower in winter than on the ice-bound straits of the Archipelago. This would give rise to southerly winds to equalize the pressure, and thus explain the greater number of these winds in Northumberland Sound and Port Refuge. They do not prevail at these places, because the depression about Iceland is still felt there as well as the depression which must exist on the open waters of Davis' Strait and Smith's Sound. As the other stations of the Archipelago, except Mclville and Dealy Island, are much nearer to Davis' Strait, they must feel its influence much more, while a great extent of islands and frozen bays and sounds separate them from the northern partly open Polar Sea.

The prevailing northerly winds in summer can be explained partly by the same cause as those of winter--the low pressure about Iceland. It is true the barometer near Iceland is not as low in summer as in winter. But in the Arctic zone of America the pressure rises also, especially from February to May; in the last-named month it is the highest of the year in most of the stations of this region.

It is probable that the pressure continues to rise in the circum-polar zone till July, thus causing the northerly winds of Arctic America. At this season air is also drawn towards the interior of North America, especially towards the region between the Rocky Mountains and 95° W. Long.

Arctic America is noted for its frequent calms in the colder part of the year—a feature observed by nearly all who wintered in these regions. They are, however, recorded in a very discordant manner in the journals of observations, showing there was a great difference in the meaning of the word "calm." This want of agreement has prevented a more elaborate discussion of this phenomenon, one of the most important in regard to the movements of the atmosphere.

Dr. Bessels has calculated the percentage of what he calls "absolute calms," for the hours when a self-registering wind-vane did not indicate any movement of air whatever, for the second winter-harbor of the U. S. Expedition, under Capt. Hall, at Polaris House or Lifeboat Cove.

I should remark, that in many of the stations the proportion of calms increases

towards March and April. In these months the cold is still intense in this region, and the pressure generally higher, so that barometric poles or areas of highest pressure are frequently met with. They are generally accompanied with calms or light winds. On the other hand, the indraught towards Iceland is less, as pressure has also risen there. (See Tables, Zones 2, 3, 4, and 5.)

In cold continental areas of lower latitudes, especially in Siberia, the greatest number of calms will be experienced in mid-winter, the time of lowest temperature and highest pressure. In March and April, when temperature is much higher, pressure decreases, and so also the number of calms.

The following figures give the percentage of winds in Greenland. Winter and summer are chosen as the two contrasting seasons of the year,1

All these stations except Sabine Island are situated on the western shore of the greatest island of the world, an island covered with large sheets of ice, and the temperature of which is much below that of the surrounding seas in winter, spring, and autumn; Smith's Sound is open the greater part of the year, though bearing large floating icebergs. Monsoon winds must be expected in these conditions, and this is really the case.

The winds of Polaris Bay have a peculiar interest, this being the most northerly station at which civilized man has ever wintered.1 Polar winds prevail largely in spring and winter. Yet there is a great difference between the N. E. and E. winds. The second prevail if the number alone is regarded, but the N. E. prevail

In all cases, except when specified, the percentages are calculated from the winds collected by Prof. Coffin.

2 From the observations of Dr. Bessels, of Capt. Hall's Expedition.

3 Observations of the Second German Polar Expedition, under Capt. Koldewey.

I owe this information on the winds of Polaris Bay and Lifeboat Cove to Dr. Bessels, who has kindly permitted the use of his observations.