seepage under 4 feet of water is insignificant. In rare cases it may be necessary to throw in manure, or at least straw or hay, and again soak and tramp, but generally there is enough organic matter in the surface soil to puddle sufficiently without the addition of any other material. Some preparations have been advertised as linings for reservoirs, but the cheapest and undoubtedly the best lining is as deep puddling as can be done with heavy animals.

Every reservoir must, of course, be provided with a suitable outlet for the water, and unless careful attention to the pumping can be relied upon there should be also an overflow or waste-way. The outlet flume should be provided when the reservoir is laid out and before the banks are completed. The bottom of the outlet should be placed as low as the bottom of the ditch outside, as shown in fig. 2, but is usually not so low as the bottom of the reservoir inside, so that the water is not all drawn from the reservoir. It is found that reservoirs retain their ability to resist seepage better if never entirely drained. This is especially true in the winter season, when, if allowed to freeze, the puddling of the bottom frequently becomes so deteriorated as to necessitate renewal. The flume should be long enough to reach entirely through the embankment. It is usually made of 2-inch planks, with a clear opening of 6 by 8 inches. Where large areas are to be irrigated, making rapidity of supply essential, the box should be made larger, and for field work an opening of 12 by 16 inches will be found ample. It is not always necessary to use the full size of the opening, but the flow is regulated by the extent to which the valve or gate is raised.

Several forms of valve have been used. An excellent one is shown in the illustration, fig. 3, from a Kansas experiment station bulletin.' This gives a perspective view of the outlet box with guides attached for holding the valve or gate. This is faced with leather securely held by tacks. The gate is best operated by a carpenter's bench-screw attached to a suitable support and engaging a nut on the upright standard. An efficient gate is also shown in the illustration, Pl. IV, operated by a lever and rod extending through the outlet box. Whatever form of flume is used, it is bet

SUPPLY PIPE

H

FIG. 2.-Diagram of reservoir and outlet.

ter to have one or two bands of cleats around it, taking care that

'Small fruits by irrigation: Experiment station, Kansas Agricultural College, Bulletin No. 55, December, 1895, p. 134.

they fit water-tight to the planks. In placing the flume it should be very thoroughly embedded in the bank, care being taken to ram the earth firmly against the planks, especially on the under side, in order to prevent the water from creeping along between the planks and the earth. The extra work in properly bedding the flume is worth many times its cost in security against leaks.

Little expense is needed to prepare an irrigation reservoir for the abode of edible fishes. It is requisite, however, to place wire screening in front of the gate, so that when water is to be drawn from the pond no fish can pass through the flume. A light framework is usually constructed, of such dimensions as to allow the gate to be opened without interfering with it, and this is covered with common wire netting, of the sort used for fly screens. This rusts out somewhat readily,

but is cheap and easily renewed. Screens of brass wire, while costing more at the outset, last much longer. The water as pumped from the wells into the reservoir is perfectly clear, and is usually at a temperature of about 58° F. This is not ideal water for the abode of carp; nevertheless, this variety has been more cultivated than any other, largely on account of its prolific increase. The growth of the carp is, however, somewhat disappointing, and this is probably due to the fact that, owing to the continual additions of the cold well water from the pump, the temperature of the water is below that at which the fish thrives best. Channel cat, crappie, and black bass have been found to do well. Some attempts have been made to grow trout in these ponds, but without success.

If irrigation reservoirs are not otherwise used in the winter, they form excellent ice ponds, and make available what would otherwise be an expensive luxury. But if water is continually pumped into the pond, especially in the more southern portions of the district, ice is but scantily formed. It is possible, however, for a farmer who has a reservoir to provide himself with a sufficient supply for the dairy and other purposes, even though he use his reservoir for irrigation during much of the winter.

DITCHES.

Not every tract of land has such conformation as renders irrigation practicable. Land the surface of which is formed into mounds of considerable magnitude may well be discarded, at least until that of more even contour shall have been brought under irrigation. Ridges and terraces are not especially objectionable features. It scarcely seems necessary to remark that irrigation is practicable only where water can be induced to flow over the land, and that water will not flow uphill, although the eye is often deceived and the novice at his first observation of the use of water through irrigation ditches receives the impression that in some parts of the work water is actually flowing up grade. And the most practiced are sometimes, indeed frequently, mistaken as to whether the surface of the land rises or falls in a given direction. It is therefore important, in the location of ditches, that a preliminary survey of the land be made; and it is a good plan to make a contour map of the land to be irrigated before locating the ditches. In irrigating from reservoirs it is essential, as before stated, that the reservoir be placed on the highest part of the land to be irrigated.

Ditches, as well as reservoirs, should be kept on land at least as high as, and generally a little higher than, that to which they are to supply the water. One of the largest and most successful irrigators of farm crops on the plains, C. D. Perry, of Englewood, Kansas, holds that in laying off the ditches it is necessary to forget all about directions of land lines and points of the compass, and be governed entirely by the contour of the surface to be irrigated (see Pl. VI).