the normal flood section, excepting for a moderate distance below the dams. Alterations in the shape of the shore lines, projecting abutments, and bridge piers, are more competent to change the channels, create bars, etc., than equal areas of obstructions laid down systematically across the river in the form of dams. This opinion is presaged on the fact that dams create only local horizontal disturbances along the axis of the streams which are soon lost, whereas obstructions creating transverse deflections may extend from shore to shore, developing secondary and tertiary effects, sometimes worse than the original cause and quite remote from it. The writer agrees with Mr. Walker in recommending wickets in lock gates for filling and emptying locks up to the limit where their capacity is sufficient with, say, 8-foot lifts, to do the work either of filling or emptying the chambers in not to exceed six minutes. It would not be advisable for several reasons to make wickets in gates large enough to fill locks the size of those on the Lower Monongahela, viz: 360 by 56 foot area, in the time specified. On the Monongahela wickets in the gates are retained, but the filling and emptying of locks is usually done by means of conduits with either cylindrical or butterfly valves placed in the walls. Four minutes' time for 8-foot lift is required. The filling is done from conduits extending down about 80 feet below the upper gates, having ten orifices, or five in each wall. The movement of steamer wheels appears to keep the lock floors clear of obstructions. The walls are not increased in section where they have conduits in them so that their use on the Monongahela, and also as proposed for the Panama, as well as on other locks, reduces the amount of concrete in construction, but not the cost, when forms are taken into account. Between the chamber faces and the conduits the intervening concrete, on Monongahela River locks, is 3 feet thick, and considered as columns the load on the wall between conduit opening is, in the maximum case, not quite 4 tons per square foot. There is no indication of structural weakness in the walls, although upon unstable foundations lines of weakness, referred to by Mr. Walker, might develop.

Also, in the case of cast-iron cylindrical valves, referred to by Mr. Walker, the writer agrees with him in recommending their merits to the consideration of engineers. They are reliable and durable, provided due attention is paid to their setting in such manner that vibrations will be reduced to a minimum. Quite recently, after several years' constant day and night service with no repairs necessary, several cylindrical valves of 8-foot diameter broke down almost simultaneously at Lock No. 3 on the Monongahela. On one ring having 38 screw-bolts 32 were loose or broken off, the standards in every case broken, etc. There was no escape from the conclusion that the trouble was due to vibration which, perhaps, were imperceptible at first. By a misunderstanding of the plan it so happened that the standards had only half the metal in them that was called for. This mistake, however, is scarcely sufficient to account for the breakage.

Regarding reaction below dams during freshets, as affected by their form or section, much might be said. At Dam No. 1, on the Monongahela River, which is being rebuilt with a concrete top, with horizontal apron below, work ceased last fall with 490 feet of new work completed, leaving 460 feet of the old timber structure (built in 1840) with lower slope with 1 foot vertical to 4 feet horizontal untouched. Soundings were taken for several hundred yards below both sections, and during the winter to date a number of photographs illustrating wave action at different stages in the river have been made. Until the work of duplicating the soundings to note difference in scour, if any, has been done, no opinions will be ventured.

Experience on the Monongahela on the whole favors the location of locks on the concave in preference to the convex side of rivers. In the case of old Lock No. 3, on a slightly concave shore (now built on a different site) to avoid trouble with ice, a deflecting dike 600 yards above was built which remedied the difficulty very effectively. As very high floods submerge the locks on the Monongahela River from 10 to 16 feet, their location appears not to disturb the general trend of the currents. It would be a difficult proposition, perhaps, if the lock walls came up even approximately to high flood elevation. Immense deposits would be made below them, even if they were on the bend side of the river. As no two floods have the same slopes, or velocities, at corresponding depths, these varying factors result occasionally in unexpected and curious effects along the river which are difficult to explain, leaving always something to learn from "further observation." At least, the writer after almost three decades' observations of the rivers about Pittsburg, is still noting "unprecedented happenings." As Tennyson said (in part) "Men may come and men may go," but for the rivers “We go on forever, upsetting the finest wrought theories of engineers.

Mr. D. M. ANDREWS
Assistant Engineer

The writer believes that the location of locks, particularly on torrential streams, such as is the Cumberland River, should, first of all, be considered from the economie side, as is done in the investigation of any other great industrial enterprise.

Both the economic lift and economic navigable depth can be approximately determined, as indicated in the writer's paper: "The Economic Improvement of the Coosa and Alabama Rivers, in Georgia and Alabama," published in Vol. L, Transactions, American Society of Civil Engineers, 1903.

The economic lift affects the ultimate cost of the project; the economic depth, its success as a commercial enterprise.

The development of power should, under present industrial conditions, be considered in the location of locks and dams on navigable streams, and those locations which best conserve the interests

of navigation and the present or future needs of the people for power are the ones that should be chosen.

In a system of improvement on a torrential stream consisting of a series of locks and dams, the crest of each dam should be so placed as to give the minimum navigable depth on the lower miter sill of the lock next above without resort to much channel excavation, for the cost of subaqueous rock excavation will soon outrun the cost of the additional height of lock and dam required o overcome it.

The writer has had experience with filling and emptying through wall culverts, and emptying through valves in the lower gates, and advocates wall culverts extending the entire length of the lock, and filling through submerged openings located at equidistant points along the chamber walls, and culverts emptying into the tail bay, for the following reasons: The filling is accomplished with less disturbance than by other methods, and the accumulation of silt above the lower gates is prevented. In silt-bearing rivers, silt is deposited against the lower gates where the emptying is done. through valves in the gates, and the writer has encountered trouble from silt with this type of fitting; on the other hand, where the bottom of the intakes of emptying culverts were placed 2 feet below the bottom of gates, no trouble from this cause was experienced.

The writer believes that the balanced, or butterfly, type of valve is best for locks under 20-foot lift, where loss from leakage need not be considered, on account of simplicity of construction and ease of operation.

Gate maneuvering gear, especially in the South where there is navigation the year round, should always be placed above water.

The writer has experienced the trouble from backlash at dams with stepped lower slopes, described by Mr. Walker, and overcame the trouble as Mr. Walker did, by the substitution of a long downstream slope. The method the writer employed is fully described in his paper on foundations, PROFESSIONAL MEMOIRS, Vol I, No. 4. In fixing the guard for the proposed locks on the Coosa River, the overfall at each dam was computed for what may be called the extreme ordinary discharge, and the top of each lock so placed as to be at the point of submergence when that discharge is reached. The guards so determined were in no case excessive.

Locks are now of standard design, which probably can not be improved upon. The writer, however, can not agree with Mr. Walker that the equipment "is no place for experiment." In the absence of an experiment station for trying out new designs, a strict adherence to standards precludes all hope for improved equipment.

Short canals should never be employed as part of a system of improvement on the silt-bearing rivers of the South and West; for, as shown by Mr. Walker, the excessive silting is too great a charge against maintenance.

Where rock is sought for foundation, the writer believes that no test should be relied on, other than the core-drill test; for he is at present confronted with the problem of founding a lock on bad

foundation tested with the steel drill and pronounced "good foundation."

test.

An inspection of the "country rock" is part of the foundation Limestone should be regarded with suspicion, and chert rejected. Chert, when in situ, represents a replacement of limestone with silica, and is almost sure to be faulty. If, however, it is the only available foundation, it should be tested with more than usual care.

Mr. THOMAS E. JEFFRIES

Assistant Engineer

Experience on the Kanawha River with fixed dams is limited to two locations. I do not not feel that I am in position to add much to what Mr. Walker has said on the subject.

I am of the opinion that the scour below fixed dams at times is underestimated, and even when a rock foundation is available the lower side of the dam should, if possible, be protected.

I have for some years advocated a slight batter on the chamber side of lock walls as a protection against damage from bolt-heads and iron bands on boats, particularly coal barges. Lock walls on this river with vertical chamber faces are badly worn and scarred, while others with battered faces, though of greater age, look nearly as good as when first built. I am aware that this is objected to by the profession as a rule, because it is thought boats may get jammed by being lowered in locking from a wider to a narrower width. But why could not the upper entrance be choked so that any boat that came in on the upper pool could get out one the lower one? Experience on this river with battered walls has been good, and though there were two boats jammed in a lock when the first ones were put in operation, it has never occurred since, and what else could you expect when an attempt was made to take two barges 25 feet wide through a lock 50 feet wide?

In drilling for rock, good results can be had by using a burley churn drill inside of a 3-inch gas pipe. The gas pipe will follow the drill to the rock and a very good idea of the quality of the rock can be formed from the drillings pumped out with a cheaply constructed sand pump. The sand pump can be made by crimping the edges of one end of a piece of gas pipe and a large marble used for a valve. While it may be a little more expensive to get up a rig for this kind of a drill the results are much better than a steel rod, and it can be used where a core drill is not available. For deep drilling a portable scaffold should be provided for the top drill man and a lever and chain for pulling the gas pipe casing. A drill of this kind was put to bed rock at Dam No. 11, 43 feet below low water.

The upper valves used in the locks with stationary dams on the Kanawha are placed horizontally in the upper bay, the water passing through culverts located under the miter sill. These valves work satisfactorily, but they do not wash the sediment from the lock

chamber, where it accumulates to an extent that requires dredging every few years.

The river banks below both of the stationary dams have been badly washed for lack of early protection. At one site it is worse below the lock, and at the other below the abutment.

Mr. W. H. MCALPINE
Assistant Engineer

I have a few remarks to make on the very interesting paper by Mr. Walker on "Elements Affecting Lock Construction on Canalized Rivers Having Fixed Dams," basing these remarks on the experience I have had in operating and care of the Kentucky River, Kentucky.

Under valves" I may mention that the cylindrical valves for filling valves have proved satisfactory on the Kentucky River, where they have been in use several years. They are easy to operate, and so far have required practically no repairs, and the mechanism shows no signs of wear.

At Lock No. 9, Kentucky River, Kentucky, with an 18-foot lift, a large single cylindrical valve in each wall is used for both filling and emptying the lock, and its use for both purposes has been entirely satisfactory. Both upper and lower valves discharge through cross culverts above the sill, and thence through ports or horizontal culverts under the sill. No more trouble has been experienced with deposit in the lock chamber and approaches than at other locks with valves in the gates.

At Locks Nos. 8, 10, 11, and 12, having an 18-foot lift, the locks are emptied by so-called "balanced valves" in the lower gates. These valves vary in size and design, some having a vertical and others a horizontal axis, but none have given satisfaction. They are hard to operate, easily get out of order, are liable to damage from drift, and are very difficult to repair, requiring unwatering with a batten at one opening and a "dry dock" at the other. Almost all of these valves which have been operated for several years have required repairs. I have yet to see a balanced valve of considerable size that operated easily under a head of 14 feet or more. Most of these valves require two stout men to operate them. The cylindrical valve for emptying is protected from drift by screens, and consequently is less liable to damage. The liability of the balance valve in the gate to injury from drift or in other ways, and its difficulty of operation, render its use unsatisfactory for either filling or emptying valves under large heads.

The very satisfactory service of the cylindrical valves for emptying lock at Lock No. 9, since the completion of the lock in 1903, would seem to justify their use for this purpose for locks of high lifts, especially where there is room to have the discharge through horizontal culverts under sill.

Referring to the author's remarks on dams, I may state that in