The quantity of water carried by the Sevier will now be considered. This, of course, is highly variable from month to month. The time for measurement, if the true irrigating capacity of the stream is to be considered, should be that time at which the ratio of water in the stream to the amount required is smallest. At different stages of growth of the crops the amount of water required differs considerably. The largest amount is needed about the time the seeds of the grain begin to fill out. Ordinarily this is in the latter part of July and early in August throughout the lower and most extensive portion of the valley, and a week later in the upper portions. At this season the water is not at its minimum. There is a gradual diminution of the flow during July, but the great shrinkage of the stream occurs during the middle of August, just after, or sometimes even during, those irrigations in which the greatest amount is required. The critical period of the crops occurs, therefore, just before, and sometimes dangerously near, the period of rapid decline in the water supply. It will therefore be evident that it is not a very easy matter to determine the exact stage of water which can serve as a criterion of the irrigating capacity. My own measurements, however, were hardly a matter of choice, but were made at the most advantageous period which could be selected without interfering with the primary objects of the expedition. The Sevier was measured at the junction of the two main forks, at the north end of Circle Valley, on the 6th and 7th of July. The method adopted was first to find a section of the water at a given point by soundings and by actual measurement of the width of the water surfaces, and measuring the surface velocity by means of floats. The most probable mean result of several measurements was found to be 410 cubic feet per second for the East Fork, and 450 feet per second for the South Fork, or a total of 860 feet. While this measurement was made the South Fork was being drawn upon above for the watering of about 1,100 acres near Panguitch, 35 miles farther up the stream, and also for watering about 600 acres in Circle Valley, about 3 to 4 miles above. The amount of water used in Circle Valley was probably greater than that at Panguitch, since the method employed was much more wasteful, and no provision made for returning the tail water to the stream. On the other hand, a large proportion of the tail water from both places finds its way back to the channel in spite of waste, but how much it is impossible to conjecture. I think, however, that 75 cubic feet per second would cover the loss from these sources.