would have the power to destroy a great deal of what had been laboriously done.

The Culebra Cut forming the low point for many square miles of territory, and coinciding for a considerable distance with the Obispo River valley, would naturally collect vast quantities of water were steps not taken to prevent it. The course of the Obispo River was artificially changed, beginning at the point where it approached the cut. The total length of the new river channel as originally built was 51 miles, from a point on the east side of the Culebra Cut, near the foot of Gold Hill, to a point clear of the cut, and finally discharging into the Chagres River. On account of slides encountered during construction work, the Obispo diversion gave way, and the flow of the river entered the cut for three days, causing inconvenience and damage. A new diversion channel was constructed with great speed. That the Obispo diversion was no small problem may be noted from the fact that in six years a total of 1,200,000 cubic yards of excavation was necessary, of which nearly 40 per cent was in rock and the total cost was over $1,000,000. The diversion was able to carry 6000 cubic feet of water per second. The Camacho diversion on the opposite side of the cut was similarly built.

These two diversions take waters which flow toward the Atlantic. The Rio Grande River formerly flowed through part of the area excavated on the Pacific side of the Continental Divide. It was similarly diverted, and a dike was constructed across the south end of the canal to prevent access of the river water. Keeping

water out of the cut also kept out the silt which would inevitably have come down with the freshets.

The elevation of the bottom of the cut was 40 feet, which was lower than the Chagres River where it joined the canal, and a dam was built across the cut with its crest at elevation 73 to prevent the river from flowing into it.

The natural streams being thus prevented from entering the work, it only remained to get rid of the water which originated along 8 miles of cut. This was done by means of centrifugal pumps at low points in the cut, which discharged the water over the dams. Excavation at a new level was always preceded by the cutting of a pioneer trench down the middle of the canal, in which all the water was collected and carried to the pumping stations. The summit during construction was at Culebra. Drainage to the south was carried to Pedro Miguel until August, 1911, when the flow was taken through the center-wall culvert of the Pedro Miguel Lock. The drainage to the north was disposed of by pumping.

The Chagres had no opportunity to interfere with the Culebra Cut, but had ample opportunity by virtue of its location, to threaten the work on the Gatun Dam. This problem was handled with ingenuity by the engineers. The portions of the dam not accessible to the river were constructed first. The spillway was built with its foundations on rock and with the river kept out by cofferdams and otherwise. In the meantime the Chagres River flowed through the west diversion built by the French. When the spillway, in 1911,

had been constructed to elevation 10 feet above sea level, and the earth dam well above this elevation, the channels of the Chagres were closed by carrying the dam across, and the water then rose in Gatun Lake until it flowed over the concrete work of the spillway during the rainy season. This depth kept the Panama Railroad, still on the old line in the Chagres valley, free of water. During the following season the railroad was transferred to the relocated or high line above the final level of Gatun Lake. The earth dam was continually kept at an elevation well above that of the concrete work of the spillway, and the next step during the dry season consisted of constructing four very large culverts in a part of the spillway temporarily protected from the flow of the water and controlled by gates. There were provisions for placing stop planks for closing the openings at some future time. When these culverts were completed, the dry-weather flow of the river was carried through them, and the remaining concrete work of the spillway progressed as long as the dry season lasted, and as long as the culverts were able to carry the flow. During the rainy season the flow was again over the concrete work of the spillway now carried to elevation 50. Proceeding thus, the spillway was completed, and the final step, when the spillway was entirely done, consisted in placing the stop planks before the entrance to the four culverts and filling them with concrete.

The Gatun Locks extend to a depth of about 55 feet below mean sea level, and the water was kept off the site by means of a temporary dam to the north of the

locks. This was built so that the excavation for the flare-walls might be done by dredges, as the material was too soft to hold steam shovels. Inasmuch as the dredges could not dig the full depth of 70 feet, a small lake was formed over the area of the flare-walls, and its elevation was lowered by pumping until the dredge could reach the bottom. This lake was kept from flowing into the partially completed locks by means of a temporary concrete dam, built between the center and side walls near the lower end of the Gatun Locks.

AIDS TO NAVIGATION

The Atlantic entrance is marked by a light of the fourth magnitude placed on the end of the Toro Point Breakwater, where there will also be a compressed-air fog whistle and a submarine bell. The tangents in the canal are defined by range lights. Vessels going in opposite directions use different ranges giving courses 250 feet apart. There are also side lights, spaced about 1 mile apart on each side of the channel. In the Culebra Cut the range lights are omitted, due to the tangents being so short and the banks too steep for placing them. There will be, instead, 35 concrete beacons at tangent points and at intermediate locations. By a system of screening, only those lights will be visible to the navigator which are necessary to define the channel where the ship may be.

The sides of the canal are further marked by acetylene, flashing gas buoys of 450 candlepower, and intermediate spar buoys.

The range lights involved much clearing of forest