place for the year 1872. (Zone 13, No. 13.) Even at stations on the Gulf coast, there are scarcely any land winds (N., N. W. and W.) observed in summer, which would be the case if there was a regular alternation of land and sea breezes.

North of Texas, throughout the whole region between 34° and 44° N. and the Rocky Mountains and Mississippi, the winds have also monsoon features, but more subdued. The prevailing winds of this region are N. and N. W. in winter and S. in summer. The cause is the same as in Arizona, New Mexico, and Texas. There are some irregularities in the mountain region (Central Colorado) but east of the mountains, in Nebraska and Iowa, the general character is again strongly marked. It is less the case in S. E. Minnesota, but even there the winds are southerly in summer, and deflected to the S. E. by the direction of the Mississippi Valley. In N. E. Arkansas and in Missouri the difference between winter and summer is still less marked. This is an approach to the character of the region between the Mississippi and the Appalachian chain, where there is no difference whatever between the seasons, the mean direction being about W. S. W. the whole year round. (See Plate 8.)

The tables for this work were printed before the results of observations on two high peaks of the Rocky Mountains could be obtained, both over 14,000 feet high. A meteorological station was established on Pike's Peak in the end of 1873, by the United States Signal Service, and the "Report for 1874" contains the means of observations for the first twelve months. I have given them in percentages, adding the station of Colorado Springs, at the eastern base of Pike's Peak. On Mount Lincoln the observations were made under Professor Hayden's geological survey of the territories, from 21st July, 1871, to the end of January, 1874. Both Peak and Mount Lincoln are situated in the central part of Colorado.

Pike's

The difference between Pike's Peak and Colorado Springs seems to give a much greater proportion of S. W. and W. winds at the higher station, and a smaller amount of N., especially in summer. This agrees with the generally entertained. opinion as to the prevailing direction of the upper atmospheric current from the W. S. W. in the middle and northern latitudes. In any case more observations are necessary in this respect.

The mean direction of the wind in the region north of Texas is:

87 July, 1875,

Here, again, as also shown by the maps (Plates 8 and 11), summer is the season which exhibits more regularity, the mean direction being everywhere between S. E. and S. W. The ratio of the resultant is greatest in the Indian Territory and Kansas, i. e., due north of the Gulf coast of Texas, and far from the influence of mountains. It is least in Missouri and N. E. Arkansas.

In winter the winds incline much more to the west than in Texas, being even S. of west, in East Missouri, N. E. Arkansas, and in S. E. Minnesota, i. e., in the extreme east of this region. Except in these regions there is a tolerably good agreement between the other stations.

The greatest difference between this region and Texas is seen in spring, as shown in Plate 8, when the winds are everywhere more or less westerly, except in the Indian Territory. Probably the cause is this: Texas being situated in a lower latitude is earlier heated, and the air from the Gulf of Mexico is sooner drawn in. The region here considered being further to the north, ascending currents are not established as early. Besides, when the lowlands between 34° to 42° N. are already heated, and an ascending current established over them, the deficiency is partly supplied by the cold air from the plateaus lying westward, partly by southerly winds from the Gulf of Mexico, and partly by winds from the polar regions. It is necessary to remember that the distribution of pressure in April and May is not the same as in midsummer. In the region here considered, pressure is lowest in May, while in Utah, and probably also on the lower Colorado, it is lowest in July. In the spring the winds coming from the Gulf of Mexico will be more westerly than in summer, because their point of attraction is more easterly in the former season than in the latter.

To recapitulate: There is an extensive region in the southwest of the United States which has a common yearly period of winds, different as are its geographical features. It includes the extreme S. E. of California, Arizona, New Mexico, Southern Utah, Texas, Arkansas, the Indian Territory, Eastern Colorado, Eastern Wyoming, Southern Dacotah, Nebraska, Iowa, Kansas, and Missouri. The winds are S. E., S., or S. W. in summer, with a great ratio of the resultant in the south, diminishing

towards the north and east. In winter the winds are mostly N. and N. W. This region is equal to more than a million square miles, or about one-third of the United States, without Alaska.

See also Maps, Plates 8, 11, and 14, which clearly show this.

To the north and northeast is a country about which it is difficult to say anything definite. It includes the larger part of Wisconsin and Minnesota, Northern Michigan, Northern Dacotah, and Manitoba.

The percentages of the winds in this region are:

The

In Northern Wisconsin the influence of Lake Superior is clearly seen. winds are N. E. in summer, or from the lake; S. W. in winter, or from the land. It must be remembered that the five great lakes never entirely freeze over, and that the difference of temperature between the air over the open water and that over the land must be great. On the Canadian shore of Lake Superior (for example, at Michipicoten) the winds are N. E. in winter and S. W. in summer. In Northern Michigan the influence of the lake is not so clearly perceived. One of the stations, Marquette, is situated on a peninsula, having the lake to the east, while others have it to the north.

Yet it seems, on the whole, as shown on Plate 8, that the winds in this belt of country bear a resemblance to the monsoon region lying to the south, especially the prevalence of south winds in summer, which is seen as far as Winnipeg (49° 52′ Lat. North).

The next region we have to consider is that between the Mississippi and the Appalachian range extending southward to the Cumberland range, and northward to Lakes Michigan and Huron, and somewhat beyond Lakes Erie and Ontario. The percentage of the winds is as follows:

In the greatest part of this region S. W. and W. winds prevail winter and summer. Looking at the isobar-chart (Plate 14) we see that at all seasons the pressure is higher in the region between the Gulf of Mexico and 35° N. L., and much lower near the lakes; hence there must be a south wind, which is converted into a S. W. by the influence of the earth's rotation. In summer and autumn the pressure is generally higher in the south Atlantic States than in the same latitude further west, and it would seem that S. E. and S. winds should be frequent from this cause. But the Appalachians do not permit an exchange of air in the lower strata, and, as the difference of pressure is but slight, S. E. winds will not often blow over the mountain-chains. In the winter-months pressure is generally higher west of the Alleghanies. Air is, so to say, heaped up by the prevailing S. W. winds. (See also Maps, Pl. 8, 11, and 14.)

The daily weather-maps of the Signal Office show that the centres of storms generally take a course nearly along the northern frontier of the United States, especially in the region of the lakes. The monthly bulletins, in which the tracks of the storm-centres are laid down, show this even more clearly. Besides this, very low barometrical minima are comparatively seldom west of the Mississippi, and the pressure generally diminishes in the centre of a storm the further it advances towards the east. The storm-tracks then lie mostly to the north of the region. we are considering now. The winds during the passage of a storm must then be S. W. and W. When the storm-track is more southerly, cold and dry N. W. winds, in the rear of the storm, will be experienced. We see that in this region the N. W. is frequently observed, especially in the winter. The storm-tracks are, however, generally more southerly in winter than in summer.

See maps, Pl. 8 and 11, and for the motion of clouds, and the velocity of the winds, Plates 1 and 13.

The different parts of this region agree very well as to mean direction of the wind and even ratio of resultant: which generally amounts to about .30, which in winter is great enough for middle latitudes. In S. W. Illinois as well as in Kentucky the winds are much more southerly in summer than in the other parts of this region; which is easily accounted for by the proximity of these States to the trans-Mississippi region, where, as was shown before, the mean direction in summer is nearly due south. As there are no mountains separating the two regions, the country on both banks of the Mississippi being generally level, we must expect a gradual merging of one into the other. It was shown above that E. Missouri and N. E. Arkansas are also transition regions between the countries east and west of the Mississippi. Another exception is Toronto. The winds here were recorded with great care, partly hourly during more than ten years, so that the difference presented cannot be explained by shortness of the period. The ratio of resultant is great only in winter, and it seems that a great part of the then prevailing N. W. are land winds. Lake Ontario is to the S. E. of Toronto For this reason we should expect S. E. winds from the lake in summer, but it seems that they do not prevail to a great extent, and that N. W. winds coming from over the colder waters of Lake Huron also reach Toronto. The motion of upper clouds at this place, as shown on Plate 1, nearly coincides with the course of the lower winds, being somewhat to the west in all seasons, the difference is greatest in spring, 61°, and least in summer, 7°.

The mean direction is more northerly in spring than in other seasons. The influence of the high pressure in the polar regions is seen in this, as also that of the lakes, covered at this season with melting ice. In the other seasons the mean direction is very nearly S. 67° W., or W. S. W. (See Plates 8 and 11.)

To explain the accordance of observations in this region among themselves, it must be remembered that it is comparatively old-settled, and the observations are numerous, especially in New York, Pennsylvania and Ohio, and some of them