sunk, the process takes time, and every lot of drift sunk as: crease the pressure on the other parts of the dike, and as a the dike would sometimes break during this process. Onhowever, the raft soon silted up and not only acted as a pr to the dike, but as an effectual closing of the chute. INSUFFICIENT PROTECTION AGAINST SCOUR tion, to su the ergy ity real and be for the dri kep for to gr sil Sh eli This is responsible for more failures than any other one The protection first used was a narrow mat placed in front dike after its construction. These mats were of insufficient even if properly placed, to protect the piles against long cont erosion, as they would be undermined along the outer edge finally slide down and away from the piles. Great care was to so sink these mats that they should lie snug against the for the dike, but this could not always be accomplished, and the have been in some cases several feet of bare bottom between the and the piles. At first no provision against scour between the ing was made, but later a form of grillage mat was woven and between the rows composing the dike. As already stated, it necessary, in order to insure proper sinking, to leave liberal s around the piles or clusters and these openings invited at Then, again, in strong currents piles or clusters could not a be driven in a perpendicular position, but would spread out ser even when good sized openings were left in the mat, openingst on the water surface appeared to be amply large, the mat not lie flat on the bottom by reason of binding against the p near the bottom, and instead of preventing scour, the grillag times increased it. Again, where drift had accumulated in in of the dike, the mats could not be given that extension upstream necessary to insure lapping over the previously sunk foot Neither could mats constructed as these were be given sufficient tension downstream to protect against erosion from the over It was found that erosion is more active just in rear of the da than in front, especially if much of a head be formed. Whe thick mass of drift is collected in front of a dike, the current deflected downward, and unless the base and rear are well tected damage to the structure at some place is inevitable, and wh the dike is once gapped, be the gap ever so small, destruction go on rapidly from both ends of the gap. At some dikes the cumulated drift had a thickness in the immediate front of 18 fr Such a thick and dense raft would cause a strong downward def W I n, and this increment to the scour would soon reduce penetration LACK OF DURABILITY OF MATERIAL It was hoped when this dike work was begun that the dike would e quickly silted up and that once covered by silt they would be Sept always wet and, therefore, be preserved from rotting, at least or a considerable time. Both of these expectations were doomed e disappointment. The dikes did not rapidly silt up, the fills being Treater above and below the dikes than at them, and even when ilted up the decay above the low-water line was not prevented. Short life of the timber used for piling was subsequently in part liminated by the substitution of cypress for the willow and cottonvood previously used. Preference had been given to willow and ottonwood on account of the plentiful supply of these timbers adacent to the works and their consequent cheapness, but when it was found that the life of such piling was not sufficiently long to nsure permanent results cypress was substituted. While this latter lasts much longer than the former, it still does not fill the requirements of standing sufficiently long to insure beyond a doubt the results desired from closure works. Such a structure as a dike in the river, even when silted up, is subject above low water to conditions that in time will impair the soundness and strength of the best of materials, being alternately exposed to submergence and to atmospheric influence, and while cypress resists this action longer than the other woods mentioned, still its life under such conditions does not exceed five years, which period is rather too short to insure good results from the dikes at all locations. Trouble from this decaying timber is, of course, one that increased as time went on. It was just beginning at this stage in the history of the dikes and was to become more important in the future. the -an able. 唧 One other fault that underlay these experiments at river conhtt traction, especially as regards Gold Dust Chute, was the attempt to Point the di compl The requi templ do too much at once. To close at one stroke this chute with its large cross section and a high-water discharge of nearly half a milli cubic feet per second was almost superhuman. Such a sudda change in the regimen of the river must necessarily cause reacting forces that were practically irresistible. As stated above, all t dike work above Cross Dike No. 4 could better have been omitted. and the result better obtained by one or two additional dike farther down the chute. The dikes, it is believed, should not have been built higher than the 10-foot stage, except near the banks, With such low dikes trouble from drift would have been much less. ened, and strength and durability would have been increased a these dikes would probably in a few years have caused deposits that would have closed the chute to about this stage. Such a result would have accomplished the main object of the work; that keeping the entire low-water discharge in the main river, and it desirable, new structures resting, as a foundation, upon these de posits could have been constructed to cause still further closing f the chute. This same fault, haste, was the cause of much loss throughout the dike work above described and also in the revetment work to be mentioned hereafter. The piles of the dikes were driven far aheal of the ability of the mat party to keep up with the work, and mat construction was pushed far ahead of the supply of stone to sit them. As a result, dikes were frequently caught by high water without mattress protection, and mats were caught by the rise floating and with no means of sinking them. The latter would probably be destroyed and the former scoured out. As another example of the same haste, the attempt seems to have been made, at least in the early stages, to carry on the work regar less of the stage of river or the presence of drift. In some cases mats were left floating after the bars or dikes along which they ga were to lie were entirely submerged. The result of all this was not only that the working parties were seriously hampered by th strong currents and large amount of drift due to these high stages. but a large amount of partially finished work was destroyed, and the cost of the improvement enormously increased. The cost of this dike work up to March, 1885, was about as fol lows: Gold Dust system, $640,000; Osceola-Bullerton system. $510,000; Plum Point system, $370,000; total, $1,520,000. This, with the $690,000 expended in the same time by revetting banks, makes a total of about $2,200,000 as the cost of the Plun C carr the a that 1 Par mate on t T Point work up to this date, and yet some work was still needed on the dike system, and the bank revetment was not over one-quarter completed. The original estimate of cost of the Plum Point work has been $736,000, it being stated, however, in making this estimate: "It is believed that such additional works, as will ultimately be required to complete and render permanent the improvement contemplated in the system at the localities specified, will not exceed the amount herewith stated as needed for actual work." That is, that the total cost of construction and maintenance of the works in Plum Point Reach would not exceed twice $736,000, or approximately one and a half million dollars. This estimate and the actual cost and result of the first works constructed make an interesting comparison and show how greatly the character and cost of work on this river had been underestimated. CONTRACTION WORKS, 1885-1892 The foregoing account of the dike work at Plum Point Reach was carried up to the end of the season of 1884. The high water of 1885 tore several gaps in the works, one each in Bullerton Cross Dike No. 1 and Plum Point Cross Dikes Nos. 5 and 6 and two gaps in the main Gold Dust dikes below the intersection with Cross Dike No. 2. These latter two gaps, with the unclosed openings in Cross Dikes Nos. 3, 4, and 5, left an unobstructed The passage through Gold Dust Chute, which was, of course, deeply scoured. th In There being no funds available, no repair work was done this season. The high water of 1886 did still further damage. An additional ong gap was made in the main Gold Dust Dike. A large part of Osceola all the Cross Dike No. 3 was torn away, and the wreckage and drift from it per tore a large gap in Cross Dike No. 4 of that system, while other ese hig damage was done to this latter dike near its outer end. The cause destre of all these breaks was deficiency in strength due to decay and to other minor damage previously done and not repaired. s about During the latter part of the season a small amount of money beerton came available, and some slight repairs were made to the outer end 20,000 of Osceola Cross Dike No. 4 and to the lower dike of the Plum Point by system. Bullerton Cross Dike No. 2 had given some evidence of of the weakness, and this dike being considered of great importance, a small additional allotment was obtained for repairing it. The mate. rial was purchased and delivered, and work was just about to begin when the dike gave way and was so severely damaged that its rep became a work far in excess of the available funds, and nothing was done. SEASON OF 1887 The principal damage done by the high water of this season was a gap in Osceola Dike No. 1. Later in the season, during low water. a brush fire was communicated to Osceola Cross Dike No. 4, and a large part of the superstructure was destroyed by fire. During the season no work was done. SEASON OF 1888 This high water also did some damage to the dikes, especially at Gold Dust, where the gaps in the three lower cross dikes we widened greatly. It is thus seen that since the stoppage of work at the end of the season of 1884, four high waters had come and each had left its marks upon the dike system, while during this time practically work of repair had been done. Gaps now existed completely through all the chutes; Gold Dust; Osceola, behind both bars; ani Bullerton; and as much damage had also been done, as we shal afterwards see, to the bank revetment, it is evident that this for: years' delay occurred at a most critical time, and had resulted in well-nigh wiping out all that had previously been done. And yet, in spite of all the damage to them, the dikes had ha large results. The survey of 1888 (Plate LIX) showed that aroun the Plum Point system the bar had grown up to above the 25-for stage out to near the end of the five upper cross dikes and that the low-water line practically followed the ends of these dikes. At Osceola Bar both the upper and lower entrances to the chut had been made out to the general bar line, and the chute behind the upper bar was closed at about the 25-foot stage, and behind the lower at the 17-foot stage. The entrance between Osceola ar Bullerton bars had narrowed a great deal and was now closed at th 3-foot stage. The chute behind Bullerton had also narrowed a goo deal and was closed at both ends at the 3-foot stage. At Gold Dust also large deposits had been made; the triang formed by the main dike and Cross Dike No. 2 had filled up fair solid to near the 20-foot stage, and below the triangle there was |