large flat 3 feet or so above low water. Elmot Bar had grown upstream, and the high timbered bar now extends well around the outer end of the main and lower cross dikes. However, these lower cross dikes had been left in 1886 with gaps in them, and as gaps had occurred in the new dike below Cross Dike No. 2, there was thus an open passage through the chute which had scoured out at places to 15 feet below low water. However, below Dike No. 5 this deep cut spread out and disappeared, and a deposit of soft mud had formed up to the 3-foot stage, so this chute also was closed at low water. Funds finally became available late in the season of 1888 and with them work was again started. Bullerton Cross Dike No. 2.-This dike had been badly shattered in 1886, just as repairs were about to begin, and no work had been done since, but as the dike was considered of importance in closing the chute, it was decided to close the gap in its middle. This was at done late this season by driving a three-row section behind the #main work and connecting it with the uninjured ends. The length of this section was 300 feet. To prevent, if possible, the reopening of the gap, a method of sinking the drift in front, by means of dh pumping sand upon it, was tried with a suction dredge borrowed for that purpose. Some good was accomplished, although, owing to the the many voids in the drift mass, much of the material pumped was he carried through to be deposited below. Before this experiment could be concluded the borrowed dredge had to be returned. di Gold Dust Chute.-The system of dikes built at the head of this chute has resulted, as stated above, in closing the chute at low water, but it being considered desirable to still further close the chute, two new dikes were planned. These were to be located in the two branches into which Gold Dust Chute divides at its lower end; to these branches being known, respectively, as Elmot Chute and the hute Chute of Island 30. (Plate LX.) rrowed and Work was begun late in the season of 1888 and consisted in conOstructing the bank protection along the shores against which the wel dikes were to abut and in building the dike foot mats entirely across both chutes. The bank protections were 500 feet long on the Tennessee bank of Chute 30 and 275 feet long on the other banks of the chute and on both banks of Elmot Chute. The reason for making the Tennessee revetment so much longer than the others there was on account of deep water and extensively caving bank there. the lled The subaqueous mats were 100 feet wide, of the woven type, with Pile driving did not begin until 1890, as the contractor for the piling could not get them to the river bank until the river was ne flood stage. This stage came late in January, and piles were at once delivered. In all essential features these two dikes were copies of the Bulle ton Cross Dike No. 2, which has been fully described. (See page 6). There was one change, and that was that the dikes were designed hold all drift, even at the highest stages; they were, therefor given an elevation in the front row of 1 foot above the highest re corded floods, each succeeding row being 1 foot less in elevatiz than the one in its front. It was expected that should the rive rise high enough to float any drift over the first row, the succeed ing rows being all lower, such drift would pass over the structure without doing any damage. Work in both chutes was started simultaneously by two partis in each chute, who worked from the banks out toward each other. Very deep water was met with toward the middle of Elmot Chute and near the Tennessee shore in the other, 50 feet being found at a number of places. No unusual difficulties were met with, althougi. G as might be expected in such deep water with a high velocity progress was slow. The most efficient pile drivers with most eff cient crews were employed in the deep water, and on some days with all hands working faithfully, but one cluster of three pi would be driven, while under ordinary conditions from twelve t sixteen piles were considered a day's work. Elmot Dike was completed early in June, 1890, and by the end c that month the remaining gap of 220 feet in the Chute of 30 Dik had been closed, with a front row consisting of clusters of trip N 山 piles. The river was high and full of drift, and immediately after the closure the accumulating drift began pushing the unsupported front row out of the perpendicular. On the following morning there was a gap of 90 feet here, which rapidly widened out to 200 feet. Soundings taken in the gap showed depths of 16 feet below low water on the line of the front row, while along the fourth row there were 57 feet below the same datum. Owing to unfavorable conditions of river, no attempt was made to close the gap until August. By that time a deep hole had been scoured out in the gap of the dike; therefore, the work designed for the closure of the gap was in the form of a salient, with its apex 75 feet above the line, thus bringing the work into comparatively shoal water. (Plate LXI.) As the original dike mat extended 75 feet above the front row, no additional mats were deemed necessary. Double sets of stringers and braces were put on this piece of work, one set as usual near the top and the other at about mid-stage. The closure of the gap was completed by the end of the month. Elmot Chute Dike remained in good condition until the end of November, 1890, when a break occurred where it crosses the deep water, about 400 feet from the inner bank of the bar. As the river was very high at the time, with much drift running, no attempt was made to close the break until the following month. The break was closed by February, 1891. A few days after completion of repairs a break occurred in this dike near the Island 30 bank, where, Sowing to heavy deposits in front, no danger had been apprehended; NO 200 feet of dike was destroyed, together with the revetment on the island shore. In March following another break occurred, also on a high fill, which widened out to 350 feet before attack ceased. The dike in the chute of Island 30 remained intact until FebWarruary, when the accumulated drift broke through, making a gap of E 270 feet wide, with both ends badly shattered. Only a short spur being of this dike remained of the Tennessee end, and as that part of the with dike from the gap to the island bank had a deep mass of silted-up high drift in its front, the river at once began its efforts to regain its with original cross section by attacking the Tennessee bank. on st No work looking toward the closure of the gaps in either dike had of been done, but as the standing parts still formed a considerable from robstruction to the flood through both chutes, and gave promise of materially aiding the Gold Dust system, it was decided to put these by sections beyond danger by sinking the masses of drift in their front. ute of This drift was from 10 to 18 feet thick, and, in order to sink it Sters @ properly, heavy mats 40 to 50 feet wide were built on top, distant from the front of the dike about 40 feet. These mats were loaded with stone until they and the drift beneath them came to a state of rest. Drift from above would then float down over the submergi part and this would again be sunk, thus forming a solid mass from the bottom up, which soon silted up, so that in a short time the surface became covered with heavy growth of willow and cottonwood In the construction of these dikes the best material obtainable was used, and great care was taken in building them. The early dikes built in this reach had been built low in exposed places, so as to allow the drift to pass over them, but they had been seriously damaged by the passing drift catching and tearing lo some of the timbers and by other masses lodging on the dikes at high water and by their weight crushing the superstructure as the water fell. To avoid such damage, the new dikes built in Elmot and Island 30 chutes were raised above the highest recorded flood, in the hope that the danger of crushing would thus be avoided, and this hope was, in a measure, realized, for no drift passed over theme lodged on top. But the gradual accumulation of a thick mass ef drift, which would in a short time become so packed as to be almost impervious, brought a great strain against the structure, which, creasing as the water rose, finally broke through the barrier. The cause of the failure of these dikes was the large accumulătion of drift, and after the breaks had occurred it was noticed that large and deep holes existed in the gaps. At the time it wa thought that the initial cause of failure was that the heavy pre sure against the piles caused them to break, and that the holes wer scoured afterwards by the rush of water through the breaks, h this theory will hardly account for a phenomenon noticed at th Elmot Dike. Here, at the time of the November break, a consider able length of this dike was pushed bodily out to place, the pil still retaining a nearly vertical position and the braces and wire still holding them in something like their original relative position Of course, this part of the dike was soon after broken to pieces an destroyed. When this phenomenon is considered and is taken in connection with some experience later with the dam built in Gol Dust Chute, it would appear that other causes probably acted is producing these ruptures. The drift, as it accumulated, and th trash caught up by it formed nearly an impervious mass. As th thickness of this increased, the hydraulic head also increased, a with the hydraulic head the pressure on the dike. Not only thi but the hydraulic head would increase greatly the current flowing under the drift raft, until the velocity would be sufficient to scour through the interstices of the sill mats, these being of the diagonally woven type used at the time. Not only this, but the hydraulie head would be sufficient possibly to cause a flow under the mat and dike through a pervious substratum of fine sand. This might be washed out and the dike be thus undermined. But whether due to one or to both of these last-named causes, it is almost certain that scour at the base or under the base was largely instrumental in the destruction of these dikes. W the t th It has been thought by some engineers not thoroughly familiar with the conditions that prompt sinking of the drift raft would have saved the dikes. This, under the circumstances, would have been a very dangerous operation, and it is far from certain that the dikes would have been saved by it. Other drift would have formed on top and, if sunk, still others on top of that, until a nearly im pervious wall would have been formed in front of the dikes. This would have caused such an hydraulic head that it is almost certain that flow would have been set up along the bottom or through some substratum, and this would have eventually caused a break. The bottom all along this chute is composed of sand and fine silt, and when thoroughly saturated, as it must be at high water, it is not able to resist a concentrated head. These dikes just described were the last pile dikes to be built in these districts; no others have been constructed nor have any repairs been made since 1892 to the dikes existing at that time. It at t was thought to have been demonstrated that such structures were 1th not suited generally to the conditions to be met with on this river, nt and their use has been abandoned. reak The cost of the work during the period 1888 to 1892 was as folpla lows: Plum Point system, repairs, $10,000; Osceola-Bullerton bra system, repairs, $5,000; Elmot and Island 30 dikes, new work, relativ $87,000. race en top As no work on the Plum Point and Osceola-Bullerton systems has and is been since done, the subsequent history and present condition m bu (1901) of these works will be stated here. obably The bar around the shore ends of the Plum Point system has conulated tinued to increase in height, and in a few years was so covered with mass a growth of willows that thousands of cords were cut for use in ineres mattress construction. The low-water line has generally followed Not the ends of the dikes, and below them, probably due to their influ 1 |