ee dikes built out from that Towhead, the two upper of these es being opposite the shallowest part of the channel. The construction of the dikes just mentioned was planned this son, but it was not until October 20 that the plant could be prered and the work could be begun. Each of the dikes was to be 000 feet long and built in four sections. Work was begun on ike No. 1, and one section had been successfully placed when the ver rose quite rapidly, and as the next section of this dike was cated across the head of a deep pool, 38 feet deep at the existing age, work on this dike was temporarily abandoned and the plant ansferred to Dike No. 3, which was begun and completed to a ɔtal length of 980 feet. Later in the season the river fell to a more avorable stage and work on Dike No. 1 was resumed, but after anther section had been placed another rise caused a cessation of vork for the season. The result of the season's work was, therefore, the construction of one and a half dikes, or half the project, the failure to complete the work being due entirely to the unfavorable river conditions. After the dikes had been placed, some filling took place above and below each, but these fills were so small as to have no appreciable effect on the channel. SEASON OF 1899 During the high water of this year Dike No. 3 was cut off from the bar against which it had abutted. This dike had been built out from a low sand bar near its foot, and at high water all this part of the bar was submerged, and the current was such as to cut away the foot of this bar and flank the dike, and when the river fell the dike was in the middle of the channel, with the best water to the left of it, where the year before the bar had been. (Plate XX.) Above Dike No. 1 the bar had built out and had covered most of this dike, but it is probable that this unfinished dike had little if anything to do with causing this change. In addition to these, other changes had taken place; the upper left side of Phillips Bar had been cut away and the channel had changed its position, and was now deflected against Mars Towhead, below Dike No. 1, causing rapid caving of this sandy bar. The river channel now passed to the left of Dike No. 3, and some distance farther down it was divided into two parts by Bixbys Towhead. On account of the complicated condition of the channel lower down, it was decided to attempt, by dikes, to restore the channel to 96 its positeɑ of the preceding season. To this end work was e This work was carried on under great difficultis of vane heme as great as 30 feet, and the current per werpd, fie linear feet of dike were built, meg, 25 for added to Dike No. 1 of the preceding year. Ay nant of san ke the bank was revetted by means of a e fawende pat and super bank paving of stone. vine poate decima. The caving of the bank was stopped darze easen and some small fills were made. The for seworks immediately after their constratze ke seemed to be decreased and the de dla falled to open, wo that the object of restoring the channel to t were increased, but the channel to the Plate XX SEASON OF 1900 Dunng the high water of this year the bar av made i an so as to cover completely the upper 18. dike. Farther down, near Dikes Nos. town-ad had been still farther cut away, an y this caving been flanked and were left in the now passed right over them. These chang die to the change of shape of Phillips Bar than to any induene of the dikes themselves. No work was done here this season, and during low water the cutting away of the foot of Mars Towhead continued, and in October the channel that flanked Dike No. 4 of 1899 was 400 feet wide and 21 feet deep. SEASON OF 1901 During the high water of this year the left side of Phillips Bar was cut away much farther downstream, and as a result the bar opposite that is, on the right side of Mars Towhead-grew still farther downstream and built up around the 1899 dikes, and the channel passed to the right of Dike No. 3 of 1898 near its location in that year, but in producing these changes it is not probable that the dikes played an important part, though they may have aided somewhat in accelerating the growth of the sand bar, all these changes being due rather to a change in conditions farther upstream and above any possil influence of these dikes. ne dikes built at this locality can not, therefore, be considered a ess, and the reason for their failure is that here the attempt was le to use them for a purpose to which they were not suited. project for 1898 was not carried out, owing to impossible river ditions. In 1899 the dikes were placed along a caving bank and e expected to stop the caving to deflect the current away from banks, to cause large deposits along them, and to cut a new nnel beyond the outer end. None of these results were accomshed, except the temporary stopping of bank caving during low ter, and this was probably due largely to the revetments placed the ends of the dikes rather than to the dikes themselves, for, as › shall see later in the consideration of bank protection, such ructures are totally inadequate to protect the caving bank in this ver, and if they are unable to protect the bank from the attack of le current, they must be much less able to change the course of is current and the location of the channel. SEASON OF 1899 Cherokee. (Plate XXI.) Cherokee is the name of a channel rossing situated about 89 miles below Cairo. At this crossing the channel, which for some miles has hugged the left bank, passes over to the right bank. Along this right bank, just at and above the crossing, there had grown up a large and wide flat separated from the main right bank by a channel narrow and deep, but closed at its upper end. Into this channel a part of the discharge of the river passed by flowing across the flat, thus diminishing the flow in the main channel. As the available depth on this crossing was not sufficient, it was decided to attempt its improvement. For this purpose a set of dikes was designed, jutting out from the main right bank and crossing the blind channel and the low flat outside of it. This was expected to choke up the blind channel, encourage the growth of the flat into a sand bar, and in this way to increase the flow in the main channel. Operations were begun in November, after the completion of the Point Pleasant work. Two dikes were built, one 1,500 feet long across the blind channel and well out on to the flat; the other, a shorter dike, 600 feet long, across the blind channel only. These dikes were completed in December, too late, of course, for their effects to become apparent that season. SEASON OF 1900 (Plate XXI.) After the high water of this year had passed away, it was found that the dike had been quite successful. The flat had grown up to and above mean low water, around the die and upstream to the higher bar above. The blind channel a partially filled up, and the discharge through it had been av stopped. The channel crossing was much better than it had the season before, but it was thought that it might be still fr improved by extending the longer dike. This was done this 690 feet being added to the dike. SEASON OF 1901 (Plate XXI. After the decline of the high water this year, t effect of these dikes was very marked. The blind channel was pletely stopped up between the dikes, leaving a closed pool abcAround the dikes and beyond their ends the bar had built up above low water. The main channel opposite the dike was mai improved and was of ample navigable depth. Ashport Bar.-158 miles below Cairo, right bank. This la ity is in Plum Point Reach. A dike of the kind now under diss sion was built here in the season of 1900. (Plate XXII.) Chutes Ashport Bar is a large and high bar, or island, lying along th right bank of the river and separated from the main Arkansas shore by a high-water chute called Ashport Slough. Outside of the high bar and near its downstream end is a lower and more recent formation, separated from the bar proper by a depression, at plans nearly down to low water. Opposite Ashport Bar the main channel follows close to the left bank of the river along the revetted Bank of Ashport Bend and the dike across the head of Gold Dust Leaving this, the channel crosses over to the right bank. passing close to the foot of Ashport Bar. This channel had for some years needed dredging, and it was decided to try its improve mend by means of a dike. As at high water there was a fairly efrong enrrent through both Ashport Slough and the depression Goleads of Ashport Bar, it was thought best to start this dike on the main shore to the right of the mouth of the slough. completed is 2,170 feet long. It was not built in a straight line, but crosses the slough and then the depression perpendicularly, and then changes direction again, so as to be nearly normal to the curront in the crossing. The immediate object of this dike was to encourage the growth of the foot of Ashport Bar, to consolidate the different bars here, and thus narrow and improve Gold Dust crossing. Incidentally it was hoped that the bar opposite at the head of Elmot Bar might be partially cut away. This would move The dike as e crossing downstream, and by moving the point of attack down rther in Fletchers Bend would protect the head of the revetment that bend and do away with the necessity of its further extension ostream. An examinaiton made this year showed that this dike had been ery successful. (Plate XXII.) The bar had built up all along the ike nearly to the top of the frame, and at one place covered the ike entirely from sight for a distance of several hundred feet. In ›laces the fill had been as much as 25 feet. The contours along the outside of the bar were now very regular and the channel much improved, dredging not being necessary until the river had fallen to an unusually low stage, and then only because of the development of a gravel stratum in the bottom of the river which the current was unable to cut out. On the Elmot side of the channel the sand bar had been partially cut away. This had moved the point of attack so far down in Fletchers Bend that the bank above the head of the revetment in that bend had ceased to cave. Elmot. (Plate XXIII.) At this locality, situated a few miles below the one just described, two dikes were built in 1900. The large bar that had built up around the head of Elmot approached quite close to the Arkansas shore, which was revetted, the narrowest place being opposite a prominent salient in the revetted bank. Below this gorge the channel spread out and divided. The main channel followed down the Arkansas shore for several miles, but to the left of this there was a big flat in prolongation of the bar at the head of Elmot, and across this flat there was a widely distributed flow into a deep channel existing close to the Elmot shore. In this manner a large amount of the discharge of the river was taken from the main channel, and, consequently, where that main channel itself crossed over to the Elmot bank the crossing was shoal and otherwise bad. For the improvement of this locality two dikes were built, jutting out from the Elmot shore across the deep Elmot channel and the flat outside of it; the immediate object being to build up the flat and close the Elmot Chute and thus keep all the flow in the main channel. These dikes are located about 2,500 feet apart and are, respectively, 2,370 and 2,640 feet long. The upper one for 685 feet of its length was built on the dry sand bar, being carried across the bar to the high bank. This part of the dike was constructed of the same general form as the rest of the dike. First a sill mat 35 feet |