the dam, and pumping a measured amount of water into them, and noting the loss and rate of flow under fixed pressures. The natural flow of the ground water through the soil was also studied. Several model dams were built and experiments made to determine the slope of the water through the material of the dam, caused by the miniature lake on one side. Test pits were dug in the deposits, and the flow into the test pits was pumped out and measured, while at the same time, observations on the level of the ground water were taken in the neighborhood.

The dimensions of the cross-section of the dam were twice changed. The height of 135 feet above sea level, as originally proposed, was at first reduced to 115 feet, and finally to the adopted height of 105 feet. The surface slopes and width at the bottom were also changed.

GATUN SPILLWAY

During the rainy season the influx of water into Gatun Lake is much greater than the amount consumed, and the spillway through the Gatun Dam provides the outlet. It might have been placed anywhere on the rim of the lake and a channel to the sea constructed, but a favorable site on rock foundation was found on the line of the dam, which allowed the use of the bed of the Chagres for carrying the water to the sea. (See plan No. 2.)

The spillway consists of a concrete dam with means for overflow, and a concrete channel to lead the water away. It is a most important adjunct to Gatun Lake, for it not only will safely relieve the lake of dangerous

flood waters, but also will control the level of the water surface, thus accomplishing the storage of a part of the flood waters for use in the dry season. Its discharge capacity must be made equal to that of the greatest possible flood. To determine the amount of water, we must again seek information in the records of the New French Company and the succeeding records by the Americans. It is to be deplored that the old company obtained no record of the Chagres flood of 1879, known to be larger than any covered by subsequent records. The engineers' report states that, "The maximum momentary discharge of the Chagres River at Gatun is calculated from the measured Bohio discharge to be 182,000 cubic feet per second." This is over 200 times the minimum dry-weather flow.

An overflow type of spillway to carry off this amount of water would be over 2,000 feet long, and even so, its discharge capacity at the highest floods would not be sufficient, and the lake might rise five feet. For these reasons, the commission adopted a spillway with a crest that is semi-circular in plan and has fourteen openings cut through the upper part, closed by gates. The elevation of the bottom of the openings is at 69 feet above sea level, or 16 feet below the normal level of the lake. Each opening is about 45 feet wide. They are so wide, in fact, that the top of the spillway is really composed of a series of piers, with the openings containing the valves between them. When the gate is shut, its top is at elevation 88 feet above sea level, making the gate 19 feet in its vertical dimension. The gate may be raised 22 feet, or clear of a 90-foot

depth of water in the lake. This device for discharging water from the lake is far superior to the plain crest without gates, because the amount of water passing through may be very nicely controlled; furthermore, any increase in the depth of the water in the lake from sudden floods would tend to increase the capacity of each opening of the spillway, because the amount of water discharged through a weir is dependent upon the head or elevation of water which is acting on the weir.

When the lake is at elevation 87 a single gate will discharge 11,000 cubic feet per second, or 154,000 cubic feet per second for the lot. The maximum known flow of the Chagres is less than this amount; in fact, is only 137,500 cubic feet for any prolonged period, such as 33 hours. The momentary discharge may be much greater than this and has been determined as high as 186,000 cubic feet per second, but, in designing a spillway, the momentary maximum is not what is wanted. Should any flood occur which will exceed the capacity of 154,000 cubic feet per second, then, of course, the lake will begin to rise; but as it rises, the capacity of the spillway is increased until, with the lake at the improbable elevation of 92 feet, above which the water would flow over the miter gates and into the locks, the rate of discharge of the crest will be 222,000 cubic feet per second. In addition to this, water can be passed through the lock culverts both at Gatun and Pedro Miguel. The length of the period over which records of flow of the Chagres are available is insufficient to predict with any degree of certainty the probable maximum flood at some future time, and the commission

has again shown its wisdom in designing for a capacity which is quite far in advance of that required by recorded floods.

The gates themselves are constructed of heavy and massive steel work. They are of the so-called Stoney gate-valve type. The sliding frictional resistance of ordinary valves of this size would be very great. The Stoney pattern of valve overcomes this by using roller trains upon which the valve travels. Passing lengthwise along the dam and underneath the gates is a tunnel in which all the machinery for operating the gate valves is placed. A chain is fastened to each side of the gate and passes over a sprocket wheel on the adjoining pier, and then down through a vertical well to the machinery tunnel. A large screw is fastened to the end of the chain and passes through a worm. A heavy counterweight hangs on the lower end of this screw rod, leaving only frictional resistance to be overcome. The motor for operating the worms is located midway between the screw rods, thus applying equal lifting force to each end of the gate.

After passing over the crest the water slides over the face of the spillway, which is so designed as to fit the normal curve of the water. At the bottom the concrete work is curved so as to give the stream a horizontal direction. About 21 baffle piers are built within this area to retard the rapid flow. At the same time, the channel becomes contracted from a width of 414 feet, which is the length of the inside of the crest, to a width of 285 feet. The water is carried in a long sluiceway, lined with heavy concrete walls and paved

with a concrete floor, and is discharged at a safe distance from the dam into the old French Chagres diversion and then into the bed of the Chagres River, whence it continues to the sea. During the dry season of four months very little water will pass through the spillway, but in the rainy season varying amounts will pass. The average flow will be about 10,000 cubic feet per second, increased momentarily to almost 15 times that amount during periods of high flood. Over the tops of the piers which separate the gate openings, a bridge and roadway have been built, so that traffic may be carried the full length of the Gatun Dam.

So important to the success of the canal is this spillway that the commission's engineers constructed a model of the same for experimental purposes, one thirty-second of the size of the original. It was tested under various conditions and the facts thus gained were of value in making the final designs.

EXCAVATION OF CULEBRA CUT

The Culebra Cut is very generally and very justly considered the most important part of the canal construction work. (See plan No. 5.) The date of completion of the cut practically determined that of the whole canal. It was in charge of the Central Divison of the canal organization, which covered also the small amount of dredging and excavation within the limits of Gatun Lake. Total expenditures are a good measure of the magnitude and relative importance of the various items and, as given in the table elsewhere, show that about half the amount applied to construction work