134* Chap. XII. Appendix, Part 6. No. 1.

be as many inches higher than the edge of the ridge at the fur-row, as the half of the bed is yards wide. This may be betterunderstood by consulting Plate I., with the section, Fig. 2.

The ground so ploughed must not be harrowed much, andas little of the natural grass taken out of it as possible ; for na-tural grass stands the water better than artificial ; and the moregrass there is left in it, the less will be wanted to be sown. Pe-rennial rye-grass, the shakings of natural-grass hay, rib-grass,and the like, may be sown to produce grass at the first; andafterwards the water will raise grass of itself. The water maybe admitted (but with caution} in one year after it is done, if theland be firm, and plenty of grass-roots; but two years after isbetter; and it is best "always to do it late in the spring, as itthen grows against autumn, and is in no danger of being sweptClF with floods, even though subject to be overflowed.

The water, in floating a meadow, should be made to flow overthe beds about one inch deep, if the descent from the gutter tothe drain be one inch in a yard ; but if the descent be greater,the same quantity of water will appear rather too little in 'depth.Less water than this is not strong enough to carry a sufficientquantity of manure or sediment over the surface of the ground,and to shelter the grass ; if much more than that quantity beused, it hurts the grass, by keeping the air from it.

All the water that I have seen used for floating here, in theNorth, deposits more or less of light manure on the surface ofthe meadows, particularly in going over the ground the firsttime ; but no water is so good after it hath been used twoor three times, as when it comes first out of the river. Wa-ter should never be used above twice after it comes out ofthe stream, before it be returned into the channel again. But ifit be again returned to the channel of the river, then the waterrecruits itself, by mud that may be deposited in the bed of theriver, or by water that may pass by the meadow, which has notbefore been strained ; and if this be not the case always, it iscertainly so in rainy weather, which swells the stream, and makesit flow over the wears or dams, or over its sides, and so passesby. In floods, the water is always richest; and one good flood■will deposit as much manure on a meadow in 24 hours, as a clearstream will do in two or three weeks ; therefore autumn floodsought always to be turned over the meadows if possible. If itbe wanted to make several floated meadows in succession downthe sides of a river, the water should never be used above twice,without returning it again to the channel of the river. Thisperhaps may increase the expense, by causing more wears ordams, &c.; but the advantage of getting a part of the waterin its original .state, particularly in floods, will amply repay thedifference. I have always found that water, used more thantwice, without being again recruited as above, does but very