whole weir will take place in 15 seconds. Making due allowance for losses of pressure, we can still feel sure that each complete maneuver will take place in less than a minute, whatever be the length dammed.

The cocks can readily be turned by a float, and consequently they can at will be passive or automatic.

The system which has just been described has several remarkable properties, namely: Rapid and easy maneuvers without shock.

Great reduction in leakage, on account of the less number of spaces between shut

ters.

Possibility of application to great lifts and lengths.

But there are also inconveniences.

The principal objection to the system arises from the effect of freezing on the supplypipes, and on the cylinders of the jacks.

If we endeavor to avoid these effects, by placing the machines at a considerable depth below low water, it becomes very difficult to repair them. Besides, the smallest leakage in apparatus, which is almost constantly subject to heavy pressure, will stop its workings.

M. Girard proposed filling the pipes and the cylinders with a mixture of water and alcohol, which only freezes at a very low temperature. This water was to be collected in a waste-well when it came out of the cylinders, when the dam was lowered, and then repumped into the accumulator; but this expedient did not appear practical, and it was considered sufficient, at the Isle-Brulée weir, to place the communicating pipes 5 feet below the lowest water, after the whole apparatus had been subjected to a proofpressure of at least thirty atmospheres.

The pipes are placed in a masonry conduit prepared in the sole, and covered by a cast-iron plate.

To repair, while under pressure, a shutter which no longer works, a screen would probably be placed in front of this shutter, which would support itself on the neighboring shutters; the one to be repaired, being thus relieved, might be removed; but unfastening under water and replacing the cylinder and its pipe are rather delicate operations, and are very rarely admissible. Use will teach whether this is too great a difficulty in practice.

The absolute rigidity of each shutter has also been made an objection to this system. If a floating body strikes, it will not yield, and violent strains are produced both on the fastenings and on the organs themselves; as a matter of fact, there is the same trouble with a needle-dam.

On the other hand, I have seen a raft of timber drift against the navigable pass of the Evry dam, and break through, making the wickets swing, but causing no damage. In addition, there is reason to fear that sand may work in between the slides and the erosss-head of each piston, and may give rise to so much friction as to cause too great wear on the rubbing surfaces.

Lastly, when the operation of raising is begun, the prop makes a very acute angle with the shutter, and consequently the normal component, which alone performs useful work, is at first only a feeble fraction of the force exerted on the piston. It also follows that the component in the direction of the shutter is large, and produces friction and compression on the axle and its collars, which cannot be neglected.

The author's calculations on the moments of the developed forces are omitted.

The dimensions of the force-pump, which in case of need must act directly on the hydraulic jacks, without the use of the accumulator, depend on the time allowed for raising.

If we suppose, for example, that the weir should be raised in the space of ten minutes, and if the travel of each piston is 5 feet 7 inches, the volume generated is 4.24 cubic feet, and for the seven shutters 294 cubic feet. This is equivalent to a continuous discharge of 2.97 cubic feet per minute. The pump and the turbine which puts it in motion under a known fall of 16 inches, for example, as at the Ile-Brulée Dam, should be so arranged as to produce this discharge of 2.97 cubic feet per minute under the pressure which corresponds to the fall of 16 inches.

In addition, it is necessary that, after the fall increases, the pump and the turbine should still be able to produce this same discharge under the increased pressure which will result.

As I have said before, the navigable pass of the Ile-Brulée Dam is provided with Chanoine wickets. Its weir only is provided with Girard shutters, whose dimensions are smaller than those represented on Plate 15, but whose general arrangement is the same. Each shutter is 114 feet wide and 64 feet long, measured along its slope and in a plane normal to the axis of rotation. The total batter of a shutter when up is 16 inches, which corresponds to an angle with the vertical of nearly 120.

The weir is 82 feet long and has seven shutters, each interval being 13 inches.

Two of these seven shutters have Papillon valves in their chases.

The upper pool is at the reference 314.72 above the sea; the lower pool is at the reference 308.65; whence results a fall of 6.07 feet.

The crown of the masonry which forms the permanent part is at the reference 308.15; that is, at half a foot below the level of the lower pool. The axis of rotation fastened on this crown is at 3 inches below the level of the lower pool.

A shutter, with all the iron-work, weighs 2,552 pounds.

A piston, with its cross-head, weighs 3,938 pounds.

It is conceded that the pressure in the reservoir should not exceed 25 atmospheres. The administration, contrary to its general practice, contracted with the inventor, M. Girard, for the shutters of the Ile-Brulée weir.

The sum allowed for furnishing and setting the movable apparatus and the dependent mechanisms was 44,000 francs, ($8,360.) This would be 537 francs ($102) per running foot.

In concluding this description it should be stated that M. Cambuzat, in his report on the navigation of the Yonne and the Upper Seine, from which we have already quoted freely, states, in regard to the Girard shutters, that they work well, but that they are costly, and that a needle-dam is preferable.

As we here end our quotations from foreign reports, it is proper to state that all the measures and calculations which have been quoted from the French are given in the originals in terms of the French metrical system. They have all been transformed into English measures, as it was believed that the usefulness of this report would have been seriously impaired if a unit of measure had been used which did not immediately convey definite ideas of dimension to American engineers.

SUMMARY AND CONCLUSIONS OF THE BOARD.

We have thus far given drawings and descriptions of all the methods of which we could learn that have been used up to the present time for the management of hydraulic gates or movable dams. The most serious difficulty in applying any of these systems to the Ohio River arises from the severity of our winters and the great masses of ice that must be provided for. In order to permit these heavy masses to pass freely, it is indispensable that in the plan adopted there should be no projecting mechanism that could possibly be injured when the pass is open; in other words, the sides and bottom of the pass should then be as smooth as those of an open cut. An approximation to this condition is an absolute necessity.

Leaving temporarily out of consideration the question of the lengths or heights of passes and weirs, we will first discuss the subject of apparatus.

The following are the methods thus far described:

1. "The bear-trap," Plate 1, Figs. 1, 2, 3.

2. The wicket used on the Riom, Plate 1, Figs. 4, 5, 6, 7.

3. Thénard shutters, Plate 1, Figs. 8, 9, 10, 11, 12.

4. Thénard shutters, as modified by Fouracres, Plate 4, Figs. 30, 31, 32, 33.

5. Poirée needle-dam, Plate 2, Figs. 13, 14, 15, 16, 17, 18.

6. Combination of Poirée dam and Thénard shutters, Plate 3, Figs. 25, 26, 27, 28. 20.

7. Chanoine wickets, Plates 5, 6, 7, Figs. 37, 38, 39, 40, 41, 42, 43, 44, 45.

8. Desfontaines wickets, Plates 8, 9, Figs. 46, 47, 48, 49, 50, 51.

9. Modified Poirée needle-dam, Plate 10, Figs. 54, 55, 56, 57, 58.

10. Cuvinot drum-wickets, Plate 11, Figs. 63, 64, 65, 66.

11. Krantz wickets with ponton, Plate 12, Figs, 67, 68, 69, 79.

12. Carro gates-improved "bear-trap"-Plate 13, Figs. 71, 72, 73. 13. Girard shutters, Plate 15, Figs. 80, 81, 82, 83, 84, 85, 86.

The objections to the "bear-trap" gates have been given at length by M. De Lagrené, and need not be repeated here. The fact that this system has been in use in France for many years to provide artificial

waves for lumbering, and yet has not been adopted on the larger rivers, is sufficient to condemn it. If such a combination were used at all, it would probably take the form of the Carro modification, and therefore we need not spend time on the prototype.

Mr. H. Werner, C. E., proposed a number of "bear-trap" gates of ordinary width, giving each pair separate inlet and outlet pipes. As these gates were to be raised in succession, it was necessary to retain the water under each pair by closing each end of the space under them. For this purpose he designed solid frames of triangular shape, which were to be lifted by the main gates, rising at right angles to them, and closing each end of the space under them. The joints between the upper main gates were to be closed by narrow aprons, resting on and raised by them.

This method of construction was tested on a large model, but did not give satisfactory results. There was no trouble in raising the gates, but they could not be lowered except by hand. The difficulty seemed to arise from the side gates refusing to fall after the water was let out. The friction between them and the main gates was such as to hold them fast. This might have been expected, since even in the ordinary "beartrap" there is difficulty in lowering the gates unless the angle between them is very obtuse.

The second method, that in use on the Riom, is manifestly only designed for small lifts and for rivers not incumbered by ice or drift. It is evidently so much inferior to some of the others that no time need be spent in discussing it.

The third, fourth, and sixth methods, two of which differ only in minor details, the third being a combination of the shutter with another system, may be discussed together. The great difficulty of raising high Thénard shutters has been mentioned in the previous discussions, and besides this, the first two methods, or the Thénard shutters proper, seem to be very objectionable in a river which carries much ice or drift. The third method, which uses but one shutter, and has a Poirée dam in addition, seems to have a superfluity of apparatus. The only apparent reason for the combination is to secure rapidity of opening by using the dam as a screen behind which to raise the shutters, and subsequently, after the dam has been removed, to drop the shutters by a tripping-rod. In view of the various devices already described for promptly opening a needle-dam, it seems unnecessary to have anything else where it is used, and therefore we are of the opinion that all the methods using the ordinary Thénard shutters should be rejected.

Of the systems that are left for examination, those of Poirée and Chanoine are in very general use. Desfontaines drum wickets seem to be used only on weirs in the Marne. Girard shutters are used on the weir of a single dam on the Yonne; the Krantz system is in process of trial; and the Cuvinot and Carro systems exist only on paper.

The first question to be settled is the general arrangement of the locks and dains that should be recommended for use on the Ohio. The French system has a pass whose sole is level with the natural bottom of the river, and a weir whose permanent part is somewhat higher. When the pass is open, and the weir-wickets are down, the river is almost in its natural condition, but there will be an increase of velocity through the pass over the velocity that existed before the construction of the dam, the amount of which will depend upon the height and arrangement of the weir. The system which the board had in view was to build ordinary slack-water dams, and to make openings in these, connected with an inclined plane below, so that coal-fleets could be passed

through the openings and down the inclined pass into the lower pool. This system was naturally suggested by the existing slack-water system on the Monongahela.

Comparing it with the French systems, assuming equal lifts at the dams, we find that the difference is that, on our plan, the weir is permanent up to the level of the pool, and the sole of the navigation-pass is about half way between the low-water line and the surface of the pool. Both weir and pass are therefore higher than is the practice in France.

Assuming, for the present, that the general combination indicated is an advantageous one, we will give a description of the various styles of gates that were examined, and, after describing the one which in our judgment seemed most suitable, will resume the comparison with the French methods.

Assuming a dam with a lift in low-water of 8 feet, as is the case with the lower dams on the Monongahela, we proposed to cut down a portion of this for a width of 200 feet, more or less, and a depth of 4 feet, and to provide the passage thus made with easily-worked apparatus for opening and closing at will. We also proposed constructing below this cut an inclined plane of slight declivity, which would form a pass into the lower pool. It should be stated that when we commenced investigating this subject we were ignorant of the great progress that had been made in it by modern French engineers, as almost the only record of their work is the Annales des Ponts et Chaussées, a publication which is seldom to be found in American libraries. The copies from which the translations were made were borrowed from the library of the Headquarters of the Corps of Engineers. M. De Lagrené's work is just published, and was only received during the present month.

The following plans were especially examined by the board, the majority of them being tested by models:

Mr. S. Petitdidier, C. E., proposed the system shown in section and elevation in Plate 16, Fig. 87. It consisted of a solid dam of wood, three feet wide and six feet deep, thoroughly bolted and stiffened with iron, which moved up and down, in a narrow chamber with vertical sides, under the action of heavy counter-weights at each extremity. Friction-rollers on each side of the vertical chamber facilitated the movement. The counter-weights were hung at such a height that when the water rose they were partly submerged, and their loss of weight caused the dam to descend into the chamber, and to gradually open the pass. The counter-weights were protected from floods by heavy crib-work, and auto. matic signals were designed to indicate, both by day and night, whether the pass was entirely or only partly open. The designer claimed that this system would work automatically. The model did not sustain his views, on account of the friction of all the parts, and the whole system was open to the objection that it would only work in floods, and could not be maneuvered without complicated mechanism and a considerable expenditure of work, at a time when there might be water enough to let fleets through, and yet not sufficient to permit the pass to remain open for any length of time.

Captain J. A. Wood, of Pittsburgh, suggested a modification of the "bear-trap." He proposed to hinge the two gates together at top, and to maneuver them by chains wound around an axle under the gates, to which motion should be given by steam or water power. The upper gate turned on an axle fastened to the floor of the pass, and the axis of the lower gate was fastened to an apron at its foot, which could move horizontally in grooves on the platform of the pass. This apron was so

fastened to the platform that it could not rise. By drawing it up-stream it caused the gates to rise, and they were lowered by reversing the motion until the two gates were laid down on the platform, and the passage was left open. The gates could not be lowered into a horizontal position, as in that case it would be impossible to raise them, since any force applied to the apron would be directly transmitted to the axis of the upper gate.

No arrangement was made for admitting water under the gates, and consequently the motive-power had to lift the entire body of water resting on the gates. On this account the greatest power would be needed at the commencement of maneuvering, but at this time the lever-arm of the force would be least, and it would be acting at the greatest disadvantage. For these reasons the board did not deem it necessary to experiment on Captain Wood's system, but contented themselves with an examination of his model.

There is a remarkable similarity between the gates of Captain Wood and those designed by M. Carro. In both cases the twin gates are hinged to each other, and in each case the lower end of the lower gate is kept down on an iron track. Captain Wood, however, uses no links, and entirely overlooks the advantage of utilizing the natural waterpressure to lift his gates. On this account he would require such enormous extraneous force that his system would be impracticable for any considerable width of opening.

The Hon. F. R. Brunot, of Pittsburgh, exhibited to the board a small model of a floating hydraulic gate, which seems to meet the requirements of the case. Mr. Brunot only presented the model, leaving the completion of the details necessary to put it in practice to be elaborated by us. His system is shown, in section, on Plate 17, Figs. 90, 91, 92. It consists substantially of a hollow caisson or ponton of the length of the desired opening, (see Fig. 92,) and of suitable width and depth. A chamber is excavated in the dam at the place chosen for the gate, and when the latter is in place and lowered, the top of the caisson is even with the floor of the pass, and the passage is free. The up-stream edge of the top of the gate is securely hinged to the up-stream edge of the gate-chamber.

Two methods of maneuvering the gate were proposed by the inventor. The first method was to make a connection between the chamber and the pool above the dam, so that the hydrostatic pressure of the upper level might raise the gate. It could only rise in an inclined position, as the upper edge would be held down by the hinges. To lower the gate the connection with the upper pool would be closed and that with the lower one opened, and the gate would fall on the removal of the waterpressure. The service of the gate would be simple and inexpensive, as one man only would be required, and his work would be limited to opening and closing valves. One objection to this method of working is that a certain amount of play is necessary between the gate and the lower side of the chamber in order to prevent danger from jamming. The width of this opening could hardly be less than 2 inches, and this width into the length of the opening (200 feet) would give a total opening of 33 square feet. In order, therefore, to retain within the chamber the pressure of the upper pool, it would be necessary to have a channel of communication with it of a greater sectional area than 333 square feet. As this pressure would have to be kept up during the season of low water, there would ensue such an expenditure of water that, in many localities, the gate could not be used. On the Mononga