This section is from the book "The Potato: A Compilation of Information from Every Available Source", by Eugene H. Grubb, W. S. Guilford. Also available from Amazon: The Potato: A Compilation Of Information From Every Available Source.
It is impossible to drain too deep for the majority of farm crops. Our upland soils remain moist where the water table is 100 feet or more below the surface. It is true that a drained peat unless compacted with a heavy roller will dry out almost completely to a depth of from 3 to 6 inches. However, below that depth, it is unusual if the peat does not remain moist regardless of the depth of the water table.
When plants 'dry up' in a peat soil it is an indication that at some time the water table has been so near the surface that only a shallow root system was developed. With deep drainage from the beginning, conditions favor the development of a root system deep enough to reach through the dry layer at the surface into the moist soil below. This is the moisture that must be relied upon during a drought. It is best to prepare for a drought by deep drainage during the wet season of the year.
Tile laid deep seem to begin their increased discharge as soon after a rain as do those laid more shallow. However, the rate of increase does not seem to be so great in the deep tile in retentive clay soils. To this is probably due the fact that deep drains continue their discharge when shallow drains have become nearly or wholly dry. Furthermore, deep drains are effective for a greater distance on each side of themselves than are shallow drains. Generally speaking, tile should be laid deeper in sand, muck, and peat than in clay.
In narrow wet ravines and in springy spots the demands made upon single lines of tile may require a diameter of five or six inches. With the vast majority of laterals, particularly those in the gridiron systems, this is larger than the diameter actually required if the tile remain entirely unobstructed. However, there should be made an allowance: (1) in sandy soils, for the sand that finds its way into the tile while they are being laid, and before they are properly blinded; (2) in peat soils, for the uneven settling that may take place and cause a tile to 'jog' slightly past the adjacent one; and (3) in all soils, for the empty space that should be maintained in the top half of the tile to facilitate the entrance of water. Laterals are most efficient as collecting drains when they are less than half full. They lose this efficiency when made to discharge under a pressure head. For these reasons three-inch tile should be rarely laid except in retentive clay soils and then in lines less than five hundred feet long. Four inches is the diameter most commonly used.
Areas requiring drainage are usually so nearly level that it is necessary to make the best possible use of every inch of available fall. The gradient of laterals may be increased by laying them shallow at the head and deep at the outlet, sometimes only two feet deep at the head and three and a half at the outlet. The gradient of a main cannot be controlled in this way, because at the head it must be deep enough to receive the discharge from the laterals. However, the diameter of the main may be increased to give it the required capacity. A gradient of a tenth of a foot, or about one and a quarter inches, to a hundred feet is generally considered a minimum gradient for laterals. The gradient in large mains may be decreased to almost nothing, because the diameter of the tile itself may be considered as constituting a gradient.
With the cost of the main seldom more than $4 an acre, and the cost of the labor and tile for laterals fairly constant at about 75 cents a rod, it is the frequency of the laterals that is the chief factor in determining the cost of tile drainage. The cost will range from $20 to $35 an acre according as the laterals are placed four rods apart or eight rods apart. This frequency should be based upon principles which have been discussed in the preceding pages. It is safe to say, however, that while eight rods is the interval which in some soils may yield the greatest dividend upon the money invested in the improvement, yet there is no danger of eventually getting the laterals too close together.
Where under-drainage is desired tile are cheaper than open ditches of the same depth. Furthermore, they offer no obstacle to cultivation, take up no surface space, and are more permanent than the ditches. There are tile in Wisconsin that are working as good or better than ever at the end of thirty years, and in some states the life of tile has been much longer than that. If tile are laid more than two feet deep, it does not hurt them to freeze if they are empty when they freeze. There should be no sags in a grade line to prevent the tile from emptying themselves. Silt also has a tendency to settle in such sags. At that depth the expansion of the water in the walls of the tile has but little apparent effect upon the tile. Tile exposed to the frequent and sudden freezing and thawing occurring at an exposed outlet or on the surface of the ground may crumble in a single winter.
There is danger in sandy subsoils of the entrance of soil particles larger than can be removed by the running water. Sand enters with ease, but is carried away with difficulty. Tile laid in sandy subsoil should be 'blinded' or entirely surrounded by and packed with clay, muck, or old grass. This being done, cracks an eighth of an inch wide are permissible.
A single tile may break some years after being laid, and the earth that enters is apt to render the line useless above that point. Such a place can usually be found with ease. The broken tile should be replaced with new ones, after the earth has been removed from the tile that have been wholly or partly filled.
On springy areas having a slope sufficiently great that water runs rapidly in shallow ditches the flow of water will invariably indicate the proper direction for main and laterals. On areas less favored with a fall a few preliminary readings with a level are of value to indicate how small the fall is. If less than two feet in eighty rods, none but a surveyor's instrument should be used for the remainder of the work. If greater than that amount, a carpenter's level carefully used or perhaps the water's surface in the bottom of the trench will do. Borings made with an auger or post hole digger will show the nature of the subsoil, and the size, depth, and frequency of the laterals should be based thereon."