The Smithsonian Institution conducted for a number of years an extensive system of measurements of rainfall in the United States, and at the same time diligently collected pluvial records from every possible source. The accumulated data thus collected were placed in the hands of Mr. Charles A. Schott for reduction and discussion, and he prepared the “Smithsonian Tables of Precipitation in Rain and Snow”, which appeared in 1868. Since that time much additional material has been acquired by the continuation of the work to the present time, and also by a great increase in the number of observation stations, and so valuable is this new material that it has been determined to recompile the tables and issue a second edition. By the time the present report was called for, the preliminary computations for the tables had developed an important body of facts bearing on the climate of the Arid Region, and through the courtesy of Prof. Joseph Henry, Secretary of the Smithsonian Institution, and of Mr. Schott, they were placed at my disposal. Mr. Schott also made such a change in the order of computation as to give precedence to the states and territories which form the subject of this investigation, and by this timely favor made it possible to base the following discussion on the very latest determinations of rainfall.

The results thus made available exhibit the mean precipitation at each station of observation west of the Mississippi River for each month, for each season, and for the year. A number of other data are also tabulated, including the latitude, longitude, and altitude of each station, and the extent of each series of observations in years and months. In selecting material for the present purpose the shorter records were ignored. The variations from year to year are so great that an isolated record of a single year is of no value as an indication of the average rainfall. The mean of two or three years is almost equally liable to mislead, and only a long series of observations can afford accurate results. In the following tables no stations are included (with one exception) which show records of less than five years’ extent.

Table I shows the precipitation of the Sub-humid Region; Table II, of the Arid Region; Table III, of the San Francisco Region; and Table IV, of the Region of the Lower Columbia. The limits of each region have been given in a former chapter, and need not be repeated. In each table the first column contains the names of the stations of observation; the second, their latitudes; the third, their longitudes (west from Greenwich); and the fourth, their altitudes in feet above the level of the sea. The next four columns show for each season of the year the mean observed rainfall in inches, and their sum appears in the following column as the mean yearly rainfall. In the last column the extent of each series of observations is given in years and months. In Table I the stations are arranged by latitudes, in Tables II, III, and IV, alphabetically.

Table I.—Precipitation of Sub-humid Region.

Table II.—Precipitation of the Arid Region.

Table III.—Precipitation of the San Francisco Region.

Table IV.—Precipitation of the Region of the Lower Columbia.

DISTRIBUTION OF RAIN THROUGH THE YEAR.

In a general way the limit of agriculture without irrigation, or “dry farming”, is indicated by the curve of 20 inches rainfall, and where the rainfall is equally distributed through the year this limitation is without exception. But in certain districts the rainfall is concentrated in certain months so as to produce a “rainy season”, and wherever the temperature of the rainy season is adapted to the raising of crops it is found that “dry farming” can be carried on with less than 20 inches of annual rain. There are two such districts upon the borders of the Arid Region, and within its limits there may be a third.

First District.—Along the eastern border of the Arid Region a contrast has been observed between the results obtained at the north and at the south. In Texas 20 inches of rain are not sufficient for agriculture, while in Dakota and Minnesota a less amount is sufficient. The explanation is clearly developed by a comparison of the tables of rainfall with reference to the distribution of rain in different seasons.

Table V.—Precipitation of Texas.

Table VI.—Precipitation of Dakota.

Table V includes every station in Texas that has a record of five years or more, in all twenty stations. If the means of rainfall for the state be compared with the means for single stations, it will be seen that there is a general correspondence in the ratios pertaining to the different seasons, so that the former can fairly be considered to represent for the state the distribution through the year. Table VI presents the data for Dakota in the same way, and the correspondence between the general mean and the station mean is here exceedingly close. At each of the nine stations, the greatest rainfall is recorded in summer, the next greatest in spring, and the least in winter. Placing the two series of results in the form of percentages, they show a decided contrast:

In Dakota a rainy season is well marked, and 73 per cent. of the rain falls in spring and summer, or at the time when it is most needed by the farmer. In Texas only 52 per cent. of the rain falls in the season of agriculture. The availability of rain in the two regions is therefore in the ratio of 73 to 52, and for agricultural purposes 20 inches of rainfall in Texas is equivalent to about 15 inches in Dakota.

For the further exhibition of the subject, Table VII has been prepared, comprising stations in the Region of the Plains all the way from our northern to our southern boundary. By way of restricting attention to the practical problem of the limit of “dry farming”, only those stations are admitted which exhibit a mean annual rainfall of more than 15 and less than 25 inches. The order of arrangement is by latitudes, and in the columns at the right the seasonal rainfalls are expressed in percentages of the yearly. The column at the extreme right gives the sum of the spring and summer quotas, and is taken to express the availability of the rainfall.

Table VII.—Seasonal precipitation in the Region of the Plains.

The graduation of the ratios from north to south is apparent to inspection, but is somewhat irregular. The irregularity, however, is not greater than should be anticipated from the shortness of the terms of observation at the several stations, and it disappears when the stations are combined in natural groups. Dividing the whole series into three groups, as indicated by the cross lines in Table VII, and computing weighted means of the seasonal ratios, we have—

Table VII (a).[2]