each eighty acres or an amount sufficient to cover the land two feet deep or more, or less, an approximate knowledge of the amount of water available for use is the prime essential, and it will be found that with a very little practice this can be at least roughly obtained by some of these methods. A few such measurements at high water, medium stages and low water on any stream will prove very helpful if its use for irrigation or any other purpose is contemplated. Advice or assistance in any difficulties arising in such measurements will be freely offered in reply to any letter of inquiry sent either to the State Irrigation Engineer or to the State Geological Survey, at the postoffice, University, N. D.

THE RUN-OFF OF THE STREAMS IN NORTH DAKOTA

An interesting problem in local geology or physical geography is the relation between the rainfall of any region and the run-off of its streams. This is not a problem that can be solved from studies in a single locality and a constant ratio found once for all between the depth of the rainfall and the quantity of water flowing away in the streams; the ratio varies in different regions, in different seasons, and in years of unusually large or small rainfall. Even in a single state there are large variations if regions of as great dissimilarity in topography as the extremes of North Dakota are included.

A study of the records of the stream-gauging stations that the U. S. Geological Survey maintains in the state, which records are briefly summarized in the following pages, and a study of the diagrams for graphic illustration, bring to view many interesting facts concerning this problem, a few of which may appropriately be noted here.

"The waters of the earth are taken up by the process which we call evaporation and formed into clouds, to be again precipiOf the water which

tated to earth in the form of rain or snow.

falls upon the basin of a stream, a portion is evaporated directly by the sun; another large portion is taken up by plant growth and mostly transpired in vapor; still another portion, large in winter but very small in summer, finds its way over the surface directly into the stream, forming surface or flood flows; finally, another part sinks into the ground, to replenish the great reservoir from which plants are fed and stream flows maintained during the periods of slight rainfall, for the rainfall is frequently,

for months together, much less than the combined demands of evaporation, plant growth, and stream flow. These demands are inexorable, and it is the ground storage which is called upon to supply them when rain fails to do so.

All of these ways of disposing of the rain which falls upon the earth may be classed as either evaporation or stream flow. Evaporation we make to include direct evaporation from the surface of the earth, or from water surfaces, and also the water taken up by vegetation, most of which is transpired as vapor, but a portion of which is taken permanently into the organisms of the plants. Stream flow includes the water which passes directly over the surface to the stream, and also that which is temporarily absorbed by the earth to be slowly discharged into the streams.

If the rainfall is measured at well distributed points, so that the average rainfall over the basin of any stream is known, and if the records of the quantity flowing down the stream are kept, the ratio between the evaporation and the stream flow, or rather between the total rainfall and the portion that escapes evaporation and runs off, is at once known for that season.

North Dakota falls within three principal drainage areas as follows:

The Hudson Bay drainage area. Hudson Bay receives the waters from the Red river and its tributaries, the principal of which are the Mouse, the Pembina, and the Sheyenne, and (from Minnesota) the Red Lake. This includes the eastern side of the state and portions of the northwestern and central regions.

The Devils Lake drainage area. This interior basin which has no outlet receives the small streams from the middle northern section of the state.

The Missouri River drainage area includes, roughly speaking, that half of the state cut off by a diagonal from the southeast to north west corners. The principal tributaries of the Missouri. are, on the east side the James, on the west side the Little Missouri, Knife, Heart, and Cannon Ball.

The Red river is the largest stream in the state except the Missouri. The fall of its main valley is small, only about one foot per mile. It is the boundary between North Dakota and Minnesota and during most of the year receives its main supply from the Minnesota side, where the rainfall is somewhat greater *U. S. Geological Survey, Water Supply Paper, No. 80, by Geo, W. Rafter.

and therefore the surplus running off unevaporated is much greater and where the forests protect the water upon the surface and in the soil from the wind and from evaporation, so that the residue running off is greater; they also hold back the water of the spring and of the storms, giving a more constant perennial flow.

The Red river itself rises in Otter Tail and other lakes in Minnesota and curving southwest reaches the North Dakota boundary at Wahpeton. At Fargo the Sheyenne enters it from the west, and although the area which the Sheyenne has drained is slightly greater than that of the Red, its average flow is scarcely more than half as great.

The Red Lake river comes direct from the lakes and woods in northern Minnesota, entering the Red at Grand Forks; though the area it drains is less than three-tenths the drainage area of the Red above Grand Forks, (and is not as large as that of the Sheyenne) its flow is nearly as great as that of the Red river, and in summer is much greater, so that from thirty to sixty per cent of the water below Grand Forks comes from the Red Lake river. Its flow is from ten to twenty times that of the Sheyenne, except for a brief time in the spring.

The Red river is peculiar among the streams of the United States in that it flows north, and before the ice has broken on its lower reaches the spring thaws are in force in the upper valley and all the coulees and streams bank full. Hence the ice carried down from the upper part does not find clear passage and jams and high water are likely

The high water of 1897 is the highest recorded, coming at Grand Forks within a foot of the level of the pavements of the city streets in most of the city. In 1904 so extreme a height was not reached by almost ten feet, but the duration of high water was much longer, the water being within six feet of its highest mark this season for twenty-five days.

All the streams in this area are usually deliberate in their actions on account of their small fall; there are no sudden changes in height nor in the streams fed from the forests of Minnesota are there great extremes in flow. The difference between maximum and minimum is seen to be comparatively small for any month except sometimes in the early spring. The Sheyenne has a fall of only about three feet per mile of valley. The Red river, with its much less fall would degenerate into a mere slough,

choked with vegetation, were it not for the large amount of water that it carries. This is the case with the James, whose fall is about two feet per mile, and in a dry season its summer flow is only ten or twenty second-feet or less, insufficient to keep its channel clear.

The Mouse river, whose fall for the sixty miles of valley between Minot and Towner (through which the river meanders so as perhaps to double the total distance) is only eighty feet, is also extremely sluggish. Thus even in the unprecedented flood of April, 1904, when the stream was out of its banks and covering the whole valley bottom, it flowed away so slowly that for thirty days at Minot it was within three feet of its highest mark.

The Little Muddy, a stream flowing into the Missouri from the north at Williston, exhibits similar features, though its fall is greater.

On the other hand, the streams in the southwestern portion of the state west of the Missouri, such as the Little Missouri, Cannon Ball, Knife, and Heart, have a larger fall, from three to six feet per mile of valley in the lower portions and much more near the headwaters. This is why they have been able to erode their valleys rapidly and deeply so that cut-banks and bad lands abound. Nor are there forests anywhere in their drainage basins to retard the run-off; in the summer after a rainstorm the water runs violently down the stream, causing a sudden flood, and the stream as quickly returns to its former low stage.

This is markedly shown in the run-off tables of these streams and even more clearly in the diagrams by the great difference between maximum and minimum flow. One of these streams may run ten, twenty, or even 100 times as much water tomorrow as today.

No records of the Missouri river in this state have yet been maintained long enongh for publication. This river from the point of view of physiography is a complex problem. Part of its volume comes from the draws, coulees and rivers of the North Dakota prairies with their violent floods at the time of the spring thaw; this causes the spring rise, which then sometimes rises to an abnormal and disproportionate height as a result of ice jams and gorges. Later in the season the snows in the mountain ranges at the headwaters of the Missouri melt, keeping the river at a high-water stage in early summer.

Another point to be noted is that a single year's record gives a very inadequate idea of the future behavior of a stream. The run-off is ultimately (though some of it is delayed by soaking into the earth and thence out from springs) the residual left from a variable rainfall after plant absorption, evaporation, and all losses have taken place. For the streams of southwestern North Dakota the total amount thus unused and running off may not amount to as much as one inch over the whole drainage basin in the course of a year. A very slight difference in the amount or in the seasonal distribution of the rainfall may cause a difference in this residual (the run-off) that is large in comparison with the total amount, and this difference cannot be easily predicted.

A thorough understanding of the behavior of even a single one of our many rivers can not be obtained without a thorough study of its whole drainage area, including the topography, the grades, the forests or vegetation, the nature and condition of the soil and surface, and the geological structure, followed by a comparison between the region's rainfall and run-off records for many years. But the pages above are a preliminary contribution to the study of the subject.

Many illustrations of the apparent inconsistencies in the behavior of the different streams may be seen in the tables following. For example, in the summer and early fall months of 1904 the average flow of the Red River at Fargo was slightly more than double its average flow for the same month in 1903; the Sheyenne for those months had double the flow of the same months in 1903; the Red Lake River, one and one-half times as much; the Pembina, more than ten times as much; but the Mouse River averaged less flow than for the same months of 1903; the Knife River less, and the little Missouri less than onethird. These apparent inconsistencies could be explained, at least in part, if space permitted here, but at first glance are surprising.

TABLES SHOWING DISCHARGE IN SECOND-FEET OF

NORTH DAKOTA RIVERS

These tables were prepared from a series of daily observations, made under the direction of the U. S. Geological Survey, at each point, and give with reasonable accuracy the quantity of water flowing by the station, maximum flow, minimum flow, and mean