The collars should be examined with equal care. Con cerning the use of these, Gisborne says:

"To one advantage which is derived from the use of "collars we have not yet adverted-the increased facility "with which free water existing in the soil can find en"trance into the conduit. The collar for a 1-inch pipe 'has a circumference of three inches. The whole space "between the collar and the pipe on each side of the "collar is open, and affords no resistance to the en66 trance of water; while at the same time the superin66 cumbent arch of the collar protects the junction of two "pipes from the intrusion of particles of soil. We con"fess to some original misgivings that a pipe resting only on an inch at each end, and lying hollow, might prove "weak and liable to fracture by weight pressing on it "from above; but the fear was illusory. Small particles "of soil trickle down the sides of every drain, and the "first flow of water will deposit them in the vacant space "between the two collars. The bottom, if at all soft, will "also swell up into any vacancy. Practically, if you re'open a drain well laid with pipes and collars, you will "find them reposing in a beautiful nidus, which, when they are carefully removed, looks exactly as if it had been "moulded for them."

66

66

66

The cost of collars should not be considered an objec tion to their use; because, without collars it would not be safe, (as it is difficult to make the orifices of two pieces come exactly opposite to each other,) to use less than 2inch tiles, while, with collars, 11-inch are sufficient for the same use, and, including the cost of collars, are hardly more expensive.

It is usual, in all works on agricultural drainage, to insert tables and formula for the guidance of those who are to determine the size of tile required to discharge the water of a certain area. The practice is not adopted here,

for the reason that all such tables are without practical va.ue. The smoothness and uniformity of the bore; the rate of fall; the depth of the drain, and consequent "head," or pressure, of the water; the different effects of different soils in retarding the flow of the water to the drain; the different degrees to which angles in the line of tile affect the flow; the degree of acceleration of the flow which is caused by greater or less additions to the stream at the junction of branch drains; and other considerations, arising at every step of the calculation, render it impossible to apply delicate mathematical rules to work which is, at best, rude and unmathematical in the extreme. In sewerage, and the water supply of towns, such tables are useful, though, even in the most perfect of these operations, engineers always make large allowances for circumstances whose influence cannot be exactly measured, but in land drainage, the ordinary rules of hydrau lics have to be considered in so many different bearings, that the computations of the books are not at all reliable. For instance, Messrs. Shedd & Edson, of Boston, have prepared a series of tables, based on Smeaton's experi ments, for the different sizes of tile, laid at different inclinations, in which they state that 1-inch tile, laid with a fall of one foot in a length of one hundred feet, will discharge 12,054.81 gallons of water in 24 hours. This is equal to a rain-fall of over 350 inches per year on an acre of land. As the average annual rain-fall in the United States is about 40 inches, at least one-half of which is removed by evaporation, it would follow, from this table, that a 1-inch pipe, with the above named fall, would serve for the drainage of about 17 acres. But the calculation is again disturbed by the fact that the rain-fall is not evenly distributed over all the days of the year,-as much as six inches having been known to fall in a single 24 hours, (amounting to about 150,000 gallons per acre,) and the removal of this water in a single day would re

quire a tile nearly five inches in diameter, laid at the given fall, or a 3-inch tile laid at a fall of more than 7 feet in 100 feet. But, again, so much water could not reach a drain four feet from the surface, in so short a time, and the time required would depend very much on the charac ter of the soil. Obviously, then, these tables are worthless for our purpose. Experience has fully shown that the sizes which are recommended below are ample for practical purposes, and probably the areas to be drained by the given sizes might be greatly increased, especially with ref erence to such soils as do not allow water to percolate very freely through them.

In connection with this subject, attention is called to the following extract from the Author's Report on the Drainage, which accompanies the "Third Annual Report of the Board of Commissioners of the Central Park:"

"In order to test the efficiency of the system of drainage "employed on the Park, I have caused daily observations to be taken of the amount of water discharged from the "principal drain of the Green,' and have compared it "with the amount of rain-fall. A portion of the record of "those observations is herewith presented.

66

"In the column headed 'Rain-Fall,' the amount of water falling on one acre during the entire storm, is given "in gallons. This is computed from the record of a rain 'gauge kept on the Park.

66

"Under the head of 'Discharge,' the number of gallons of water drained from one acre during 24 hours is given. "This is computed from observations taken, once a day or 66 oftener, and supposes the discharge during the entire 66 day to be the same as at the time of taking the observa tions. It is, consequently, but approximately correct:

"The tract drained by this system, though very swampy, "before being drained, is now dry enough to walk upon "almost immediately after a storm, except when underlaid "by a stratum of frozen ground."

The area drained by the main at which these gaugings were made, is about ten acres, and, in deference to the prevailing mania for large conduits, it had been laid with 6-inch sole-tile. The greatest recorded discharge in 24 hours was (August 25th,) less than 100,000 gallons from the ten acres,— -an amount of water which did not half fill the tile, but which, according to the tables referred to, would have entirely filled it.

In view of all the information that can be gathered on the subject, the following directions are given as perfectly reliable for drains four feet or more in depth, laid on a well regulated fall of even three inches in a hundred feet:

It is not pretended that these drains will immediately remove all the water of the heaviest storms, but they will always remove it fast enough for all practical purposes, and, if the pipes are securely laid, the drains will only be benefited by the occasional cleansing they will receive when running more than full." In illustration of this statement, the following is quoted from a paper communicated by Mr. Parkes to the Royal Agricultural Society of England in 1843:

66

"Mr. Thomas Hammond, of Penshurst, (Kent,) now uses no other size for the parallel drains than the inch "tile in the table (No. 5,) having commenced with No.