hela, where the first pass with hydraulic gate should be built, the flow during summer is often insufficient to supply the lockage.

Another objection to this method is that it will oftentimes be necessary to lower the gate and open the pass when the water in the lower pool is even with the top of the chamber. In this case our experiments indicated that the water pressure on top of the gate would not suffice to make it fall.

The other method suggested by Mr. Brunot is to fill the gate with water when the chute is to be opened, and to pump the water out when it is to be closed. This will, undoubtedly, secure the desired result beyond any possibility of failure, no matter what may be the difference of level between the two pools, and this is the method which we recommend. The power necessary to do the pumping can always be had from the fall at the dam, and a turbine-wheel in a well in the abutment would be the natural method of applying it.

There is a danger that sediment may enter the chamber of the Brunot gate through the opening left for play, and also that sticks and stones may get fast in it and interfere with the movement of the gate. For this reason, and in order to secure a tighter chamber, we thought that the system shown in section in Figs. 88 and 89, on Plate 16, would be better. In this the Brunot gate is hinged to the lower edge of the chamber, and a heavy wooden apron rests on it, which is hinged above the chamber to the platform of the chute. This system gives ample play to the gate, and almost entirely obviates any trouble from sediment and from floating bodies or stones rolled along by the current. It is, in fact, a combination of the Brunot and the "bear-trap" systems. There is also a very striking similarity to the main part of the Krantz system, the chief difference being in the method of fastening the apron. It is an apparent objection to the Krantz system that the cavity prepared for the lower end of the apron might be filled with sediment or obstructed by stones and gravel rolled along the bed of the stream. We finally, however, concluded to abandon this combination, because we found that by putting a shoulder on the lower edge of the Brunot gate, as shown in Figs. 90 and 91, we could retain sufficient play, and yet make a tight fit against the chamber-wall, and because in elaborating the details of filling and opening the gate in the combined system, we found many practical difficulties that do not exist in the Brunot system proper. In the combination of the two systems the apron would have to be very wide, for the same reason as in the "bear-trap" proper, and it would be open to the same objections.

Col. P. J. Schopp, C. E., suggested what he termed a "triangular caisson dam," but as it consisted substantially of a number of Brunot caissons of triangular cross-section, each provided with an independent chamber, and maneuvered in succession, we did not deem it necessary to test the plan in a model. He proposed working his gates entirely by changing the pressure of the water under them, but, as we have previously stated, our experiments on models indicated that this method would sometimes fail, and the application of the pumping system to a number of caissons would increase the difficulties of working, without corresponding advantages.

The system now in use on the Yonne and Seine has the advantage of being the gradual growth of long years of study and experiment, and will, undoubtedly, at least on the Upper Ohio, radically improve the navigation. The question before us, therefore, is whether this system is the best that can be devised for such navigation as is used on the

Ohio River, or whether we cannot obtain a method more suited to our

wants.

As can be seen from the discussion, and from the tables which have been given, the French system gives 5 feet in depth at the head of each pool, and the lifts, or differences of level between the pools, vary on the Seine from 4 feet 5 inches, at the Melun Dam, to 9 feet 10 inches at Portà-l'Anglais. The latter lift is exceptionally great, on account of the wish to avoid constructing a dam within the limits of Paris, the next greatest lift being 6 feet 1 inch, at Ablon.

At the La Madeleine Dam, which may be taken as a sample, the tops of the wickets of the pass, when up, are 8 inches lower than those of the wickets of the weir, and they are 4 inches below the surface of the upper pool. Four inches is therefore the minimum depth of water that is expected to pour over the tops of the wickets of the pass. During this stage the water of the pool below stands against the back of the pass-wickets, at a vertical height of 4 feet 4 inches. Should the water get too low to give four inches over the tops of the pass-wickets, the dam can be tightened by temporarily covering the spaces left between the wickets.

The length of the shortest pool on the Upper Seine is 3.38 miles and that of the longest seven miles. The fall of the river between the Varennes and the La Cave Dams, a distance of 18.63 miles, is 23 feet, or an average of about 15 inches per mile, while between the same dam and that at Port-à-l'Anglais, a distance of 56 miles, the fall is 60 feet, which is an average of about 13 inches to the mile. The average fall of the Ohio in the first twenty miles of its course is 17 inches per mile, but this slope rapidly changes, and in the first fifty-six miles the fall is 53.3 feet, or at the rate of about 114 inches per mile. The slopes of the two rivers, in the portions compared, are thus seen to be nearly the same, though the greatest floods in the Ohio, in the sections compared, vary from 35 to 45 feet, while in the Upper Seine they only vary from 16 to 23 feet.. As might be expected from the greater height of floods, the banks of the Ohio are much higher above low-water than those of the Seine, and therefore higher dams can be built without greater danger of overflow. One objection to all the French systems is that the mechanism consists of a great number of parts, all of which must be kept in perfect working order, a thing which is less difficult in France than in the former country because there is a long-established and well-organized and trained body of inspectors, engineers, lock and dam tenders, and assistants, whose lives are devoted to such work, and who are thoroughly capable of attending to it. The lack of such a body in the United States makes it eminently desirable that all machinery should be of the simplest possible kind, and we believe that the plan which we recommend has at least this merit.

Another and more serious objection to the French systems comes from the greater cold of our climate, and the greater danger of injury by ice. The account already given shows that the Seine dams were greatly endangered by an unusual cold of 21°. As this is a very common temperature with us, and as the thermometer is not unfrequently below zero, it is manifest that the danger mentioned would be both more frequently encountered in this country, and more dangerous when encountered, particularly in view of the higher and more sudden floods, and the greater masses of drift-wood. Therefore, with our present information, we think it would be better to test the Brunot gate in preference to adopting the French systems thus far in use. If the test is unsatisfactory, we still have them to fall back on, while if it is a success we be

lieve that we will have a simpler, stronger, and more readily handled apparatus.

Should the system which we recommend for trial be adopted, we will find ourselves provided with a navigation that differs in many particulars from that used in France. In the latter country, as soon as the natural depth in the river is less than 5 feet, the passes are closed, and all navigation in either direction is carried on through the locks. On our system, if we can get a gate, as we think we can, that can be opened in two minutes, and closed in five, it will be quite practicable to keep up an intermittent down-stream navigation through the pass throughout almost the whole year, as the opening of the gate for not to exceed ten minutes at a time, (which, allowing for diminished discharge while being opened and shut, will make the total expenditure of water about equal to a full opening of the chute for 5 minutes,) will probably not injuriously lower the level of the upper pool. We would thus have a natural down-stream navigation throughout almost the entire year, which would be an immense advantage, since our heavy products, such as coal and manufactured iron, all go down the river. To counterbalance this advantage, we would have the disadvantage of forcing all up stream navigation, except in very high water, to pass through the locks. The latter would have to be higher than the French locks, and our expenditures for masonry and timber for the dam, and inclined plane, would much exceed theirs. To counterbalance this, we have simpler constructions, less complicated mechanism, (which is both very costly, and must be carefully watched and kept in order,) and probably less expense for attendance. It is, therefore, mainly on account of the special character of our climate and of our navigation, that we recommend that a system differing from those used in France be first tested, in preference to copying what are successful in their native country, but which might not work so well here. We wish it to be specially understood, that while we have attempted to collect all available information on this subject, we do not presume to decide the question now, but limit ourselves to recommending a preliminary experiment. Having, therefore, decided in favor of an experiment upon the plan proposed by Mr. Brunot, it now remains to elaborate its details.

There are two methods of filling the gate-by opening valves in the top of the gate itself, or by opening a pipe which communicates with the interior of the gate. As the gate rises and falls around a horizontal axis, there is some difficulty in devising an apparatus to move these valves at all times which shall itself be sheltered from floating bodies. Moreover, valves in the top of the gate are liable to injury, and they weaken the gate where it ought to be strongest. For these reasons we propose to permit water to enter by a pipe under the platform of the chute, which shall connect with the interior of the gate by several branches of flexible pipe entering just below the hinges. The main pipe will be controlled by a valve worked from the abutment, as shown in Fig. 92. It is calculated that a two-foot pipe will fill the experimental gate, 100 feet in length, in 2 minutes, which is, probably, quick enough. To empty the gate a centrifugal pump is used whose suction-pipe has a flexible length to connect it with a pipe extending to the bottom of the gate. This pump will be set in motion by a turbine wheel. The necessary power to drive this pump was calculated by assuming 5 minutes as the time for the work, and taking the capacity of the experimental gate as 5,000 cubic feet, and the lift as 6 feet. We, therefore, have a quantity of work of 375,000 foot-pounds per minute, or 11.4 horse-power. The effective work of pumps is given by Bourne as ranging from 30 to

65 per cent. of the power applied to them, and, therefore, with an assumed efficiency of 38 per cent. we find that we require 30 horse-power to work this pump.

To get the size of turbine necessary to develop this effective power, we use Francis's formula:

D= 4.85

Р h√ h

in which P = 30, the effective horse power, and

h =

6, the assumed head of water. We, therefore, find—

D= 6.93 feet.

The amount of water necessary to supply this wheel is found by the formula, also given by Mr. Francis:

Whence we find

Q = 0.5 D2 √ h

Q = 58.8 feet per second.

To supply this amount of water without great velocity, and, therefore, without sensible loss of head, the water in the channel leading to the turbine should have a velocity of not more than 3 feet per second; and, therefore, the cross-section of the channel should be about 20 square feet. We have, therefore, taken a width of 4 feet and a depth of 5. The positions of the pump and turbine are shown in Fig. 92.

It is important to have some arrangement for the automatic filling of the gate, should a sudden flood come in the night, or when the gatetender was absent or negligent. This is provided for by a stationary pipe at the far end of the gate, which is sheltered from floating bodies by the recess at the end of the chamber. The height of this pipe is such that when there is a greater depth in the chute than seven feet, the water overflows into the gate. This pipe also answers as an air-pipe during the maneuvering of the gate.

The most important navigation on the upper part of the Ohio River is the transport of coal, and as this transport is always down stream, and as the ponderous coal-fleets are not easily checked or stopped, it is very desirable that the process of lowering the gate and opening the pass should be very expeditious, while there need be no great hurry in raising the gate. The system proposed answers these ends perfectly. One man can maneuver the gate, and it can be filled with any desired rapidity. If the 2-foot pipe should not do the work fast enough, there is no difficulty whatever in using one of greater diameter. After the fleet has passed, the attendant has only to open the gate of the turbine, and in a few minutes the pass is closed. It seems hardly possible to devise a system that could promise better results.

In order to scour out any sediment that may accumulate in the chamber, a culvert is made at the far end, and inlet pipes at the abutment. As this operation would seldom be necessary, (possibly once a year,) it is believed that the management of the valve at the far end would offer no practical difficulty. Excepting this one valve, all the mechanism is on the abutment, and is therefore always accessible.

Besides the gate, the shape and details of the pass require special attention. In order to acquire knowledge in the working of passes, or chutes, as they are frequently called, the board went to Lock Haven, in Pennsylvania, and examined the chute in the dam across the Susquehanna at that place. This chute and dam are shown in Fig. 36, plate 4. The length of the chute is 1,295 feet, and its width is 31 feet. The differ

ence of level, in low water, between the pool and the river below, is 11 feet, and the slope of the chute is 1 in 142.

There is no gate at its head, as it is left open until the water above becomes too low, and then a temporary dam of plank, resting against a horizontal beam, is constructed. The chute was unfortunately closed when we visited it, but we obtained much information about it from an ex-officer of engineers, Mr. R. W. Petrikin, residing in Lock Haven. We learned that it was only used in running rafts down the river, and that there was a marked wave at its entrance, and also that rafts in passing out of the chute into the river were generally submerged. At the Williamsport chute, on the same river, rafts dive so much that in order to diminish this tendency they are forced to use, at the lower end of the chute, floating timbers with one end fastened to the bottom. Neither the wave at the head of these chutes nor the diving at the foot is injurious to rafts, but they would be to coal-fleets. A very important matter of detail is to make as much friction as possible on the bottom and sides of the chute, so as to retard the velocity of the water. The superiority of the Lock Haven chute to the others on the Susquehanna is attributed, in some measure, to the fact that it has a stone bottom, while others have wood. The stone paving has become so rough that it checks the current very appreciably.

The board believe that by a judicious widening of the head of the chute, so as to cause a large body of water to enter, the head-wave can be almost if not entirely obviated, and that a similar widening at the foot would probably work equally well there. But these and other details can only be settled by actual trial, and without such trial they would be unwilling to recommend the system for use on the Ohio. If it is a success, there need no longer be any difference of opinion about the radical improvement of that river. As such vast interests depend upon this trial, they would most urgently press its importance upon Congress. To test the whole scheme, the Monongahela Navigation Company, who need something of the kind for their own use, offer the use of the lowest dam on the Monongahela for a chute 100 feet in width, or half the width proposed for ultimate adoption on the Ohio. The two locks in their dam No. 1 are insufficient to transact the rush of business that crowds upon them whenever there is a coal-boat rise in the Ohio, and many coal-barges lose an opportunity of getting to market on account of the impossibility of getting through the locks in time. If the chute proves a success, barges can safely lie in pool No. 1 until a rise comes, and then pass down the river promptly. An indirect result would be to add pool No. 1 to the harbor of Pittsburgh, which is now greatly overcrowded.

The Monongahela Navigation Company is, therefore, equally interested with the United States in finding a successful solution to the problem before us.

They therefore, through their president, Hon. J. K. Moorhead, offer for this purpose their dain No. 1, and agree to pay one-half the cost of the experiment, it being understood that in case of success they are to become the owners of the gate and chute.

This proposition seems to us a fair one to both parties, and we would therefore recommend its acceptance. According to the most careful estimate that we have been able to make, the cost of putting in a chute 100 feet wide, with a movable hydraulic gate, the bottom of the chute being 4 feet below the crest of the dam, will be $80,000. We would therefore urgently recommend the appropriation, by Congress, of $40,000 for the purpose of experimenting with a navigable chute, to be opened