of the top should never be less than four feet, and in er posed localities it should be more. If a road will be needed around the land, it is best, if a heavy dyke is required, to make it wide enough to answer this purpose, with still wider places, at intervals, to allow vehicles to turn or to pass each other. Ordinarily, however, especially if there be c good stretch of flat meadow in front, the top of the dyke need not be more than four feet wide. Supposing such a dyke to be contemplated where the water has been known to rise two feet above the level of the meadows, requiring an embankment four feet high, it will be necessary to allow for the base a width of fourteen feet;-four feet for the width of the top, six feet for the reach of the front slope, (1 to 1,) and four feet for the reach of the back slope, (1 to 1.)

Having staked out two parallel lines, fourteen feet apart, and erected, at intervals of twenty or thirty feet, frames made of rough strips of board of the exact shape of the section of the proposed embankment, the workmen may remove the sod to a depth of six inches, laying it all on the outside of the position of the proposed embankment. The sod from the line of the ditch, from which the earth for the embankment is to be taken, should also be removed and placed with the other. This ditch should be always inside of the dyke, where it will never be exposed to the action of the sea. It should be, at the surface, broader than the base of the dyke, and five feet deep in the center, but its sides may slope from the surface of the ground directly to the center line of the bottom. This is the best form to give it, because, while it should be five feet deep, for future uses as a drain, its bottom need have no width. The great width at the surface will give such a pitch to the banks as to ensure their stability, and will yield a large amount of sod for the facing of the dyke. The edge of this ditch should be some feet away from the inner line of the embankment, leaving it a firm support or shoulder at

the original level of the ground, the sod not being removed from the interval. The next step in the work should be to throw, or wheel, the material from the ditch on to the place which has been stripped for the dyke, building it up so as to conform exactly to the profile frames, these remaining in their places, to indicate the filling necessary to make up for the settling of the material, as the water drains out of it.

As fast as a permanent shape can be given to the outer face of the dyke, it should be finished by having the sod placed against it, being laid flat wise, one on top of another, (like stone work,) in the most solid manner possible. This should be continued to the top of the slope, and the flat top of the dyke should also be sodded, the sods on the top, and on the slope, being firmly beaten to their places with the back of the spade or other suitable implement.

This will sufficiently protect the exposed parts of the work against the action of any waves that may be formed on the flat between the dyke and the deep water, while the inner slope and the banks of the ditch, not being exposed to masses of moving water, will retain their shape and will soon be covered with a new growth.* A sectional view of the above described dyke and ditch is shown in the accompanying diagram, (Fig. 47.)

The ends of the work, while the operations are suspended during spring tides, will need an extra protection of sods, but that lying out of reach of the eddies that will be formed by the receding water will not be materially affected.

In all work of this character, it is important to regulate the amount of work laid out to be done between the spring tides, to the laboring force employed, so that no un finished work will remain to be submerged and injured. When the flood comes, it should find everything finished up and protected against its ravages, so that no part of it need be done over again.

If the land is crossed by creeks, the dyke should be finished off and sodded, a little back from each bank, and when the time comes for closing the channel, sufficient force should be employed to complete the dam at a single tide, so that the returning flow shall not enter to wash away the material which has been thrown in.

If, as is often the case, these creeks are not merely tidal estuaries, but receive brooks or rivers from the upland, provision must be made, as will be hereafter directed, for either diverting the upland flow, or for allowing it to pass out at low water, through valve gates or sluices. When the dam has been made, the water behind it should never be allowed to rise to nearly the level of the full tide, and, as soon as possible, grass and willows should be grown on the bank, to add to its strength by the binding effect of their roots.

When the dyke is completed across the front of the whole flat, from the high land on one side to the high land on the other, the creeks should be closed, one after the other, commencing with the smallest, so that the experience gained in their treatment may enable the force to work more advantageously on those which carry more

water.

If the flow of water in the creek is considerable, a row of strong stakes, or piles, should be firmly driven into the bottom mud, across the whole width of the channel, at intervals of not more than one or two feet, and fascines,— bundles of brush bound together,—should be made ready on the banks, in sufficient quantity to close the spaces be

tween the piles. These will serve to prevent the washing away of the filling during construction. The pile driving, and the preparation of the fascines may be done before the closing of the channel with earth is commenced, and if upland clay or gravel, to be mixed with the local mate rial, can be economically brought to the place by boats or wagons, it will be an advantage. Everything being in readiness, a sufficient force of laborers to finish the dam in six hours should commence the work a little before dead low-water, and, (with the aid of wheelbarrows, if necessary,) throw the earth in rapidly behind the row of stakes and fascines, giving the dam sufficient width to resist the pressure of the water from without, and keeping the work always in advance of the rising of the tide, so that, during the whole operation, none of the filling shall be washed away by water flowing over its top.

If the creek has a sloping bottom, the work may be commenced earlier, as soon as the tide commences to recede, and pushed out to the center of the channel by the time the tide is out. When the dam is built, it will be best to heavily sod, or otherwise protect its surface against the action of heavy rains, which would tend to wash it away and weaken it; and the bed of the creek should be filled in back of the dam for a distance of at least fifty yards, to a height greater than that at which water will stand in the interior drains,—say to within three feet of the surface, so that there shall never be a body of water standing within that distance of the dam.

This is a necessary precaution against the attacks of muskrats, which are the principal cause of the insecurity of all salt marsh embankments. It should be a cardinal rule with all who are engaged in the construction of such works, never to allow two bodies of water, one on each side of the bank to be nearer than twenty-five yards of each other, and fifty yards would be better. Muskrats do not bore through a bank, as is often supposed, to make a pas

sage from one body of water to another, (they would find an easier road over the top); but they delight in any ele vated mound in which they can make their homes above the water level and have its entrance beneath the surface, so that their land enemies cannot invade them. When they enter for this purpose, only from one side of the dyke, they will do no harm, but if another colony is, at the same time, boring in from the other side, there is great danger that their burrows will connect, and thus form a channel for the admission of water, and destroy the work. A disregard of this requirement has caused thousands of acres of salt marsh that had been enclosed by dykes having a ditch on each side, (much the cheapest way to make them,) to be abandoned, and it has induced the invention of various costly devices for the protection of embankments against these attacks.*

When the creek or estuary to be cut off is very wide, the embankment may be carried out, at leisure, from each side, until the channel is only wide enough to allow the passage of the tide without too great a rush of water against the unfinished ends of the work; but, even in these cases, there will be economy in the use of fascines and piles from the first, or of stones if these can be readily procured. In wide streams, partial obstructions of the water

*The latest invention of this sort, is that of a series of cast iron plates, set on edge, riveted together, and driven in to such a depth as to reach from the top of the dyke to a point below low-water mark. The best that can be said of this plan is, that its adoption would do no harm. Unless the plates are driven deeply into the clay underlying the permeable soil, (and this is sometimes very deep,) they would not prevent the slight infiltration of water which could pass under them as well as through any other part of the soil, and unless the iron were very thick, the corrosive action of salt water would soon so honeycomb it that the borers would easily penetrate it; but the great objection to the use of these plates is, that they would be very costly and ineffectual. A dyke, made as described above, of the materia! of the locality, having a ditch only on the inside, and being well sodded on its outer face, would be far cheaper and better.