The estimated cost of completing the 9,035 linear feet of revetment at East Bottoms by placing 500 cubic yards of stone on the slope bank above low water is $705.50, which will make the total cost of that work $68,413.26, or an average cost of $7.5720 per linear foot.

Little Platte Bend revetment.-The protection of the bank in Little Platte Bend became imperatively necessary for the reasons stated in my last annual report. As the amount of funds available for this work (after that at East Bottoms would be completed) was small, it was found impossible to construct all the work that will eventually be required for this locality, and to use as expensive a form of protection as the Quindaro type of revetment.

After a consultation with you it was decided to protect the bend from L to C, a distance of 4,500 feet (Plate I) by a partial revetment, similar in construction to the lower 350 feet of the East Bottoms revetment.

Small piles, spaced 10 feet apart, were sunk to a depth of 12 feet by the use of a water jet fro.n one of the pile-sinkers, so as to follow approximately the contour corresponding to the stage 310 on the Kansas City gauge. Loose brush was next laid behind these piles to prevent the toe of the bank from sloughing during the hydraulic grading, and also to even up the shore-line. A woven mattress, about 88 feet wide, was next laid, extending from an elevation of about 312 (Kansas City gauge) into the stream. The transverse mattress cables were first attached to the piles and then carried to the top of the bank to their anchors.

In exceptional cases before, but usually after, the mat was woven, the bank, which in this locality on an average is about 2 feet above the ordinary June rise, was graded to a slope of about 1 on 24, at an average cost of $0.01784 per cubic yard of earth removed, including all expenses.

The grader could not work continuously, as from time to time large areas of mattress could not be sunk immediately after it was woven on account of the insufficient supply of stone, and had to be left afloat. That the cost of grading, in spite of this difficulty, is small is to be attributed to its being done after the mat had been woven, which, in connection with the loose brush placed along the water-line, prevented the toe of the bank from sloughing. Hydraulic grader No. 8 was used for this work.

The graded bank between the stages of about 309 and 323 (Kansas City gauge) was covered with stones, both where it was protected by mattress and where not so protected. As the stage of water at the time the work was done was about 311, the area of bank between this stage and 309 had to be covered under water, and therefore this part of the work had to be done at random to some extent. The stone in different places consisted either of quarry spalls or of large stone covered with spalls, or of large stone only. This variation was made to determine, if possible, which of these would afford the most effective protection to that part of the bank which was not covered by mattress.

The sinking of piles began September 6 and was completed October 2, the work being suspended for several days on account of a rise in the river. The greatest number of piles sunk in any one day was 49. After the mattress was woven the piles were sawed off even with the top of it. Three parties were employed in weaving mattress, which work was begun September 17 and completed October 20. The covering of the bank with stone was completed November 5.

Repairs to Kaw Bend Revetment.-During the spring rise in March, 1887, about 350 linear feet of bank mattress in Upper Kaw Bend, from T to S (see Plate IV), which had been constructed in September, 1885, was undermined. The cause of this damage was supposed at the time to be due to water, which had accumulated on the low ground back from the bank seeping through the bank as the stage of water in the river decreased.

This damage was repaired the following April.

During the June rise of 1887 the revetment was again undermined in places between R and W. Although the action of seep-water has no doubt had some effect, as supposed, the principal cause of the trouble appears to be as follows: It will be seen from Plate IV that the thread of the current impinges at T, where the water-way has decreased in width from about 1,000 feet in October, 1885, to 385 feet in December, 1887, the stage of water, viz, 309, on the K. C. gauge being the same for both dates. During the higher-water stages of the river the point of impact shifts from T towards W, and this part of the revetment is exposed not only to the direct impact of the current, but also to the undermining effect of strong eddies produced by the irregular outline of the bank. These forces, in connection with the heavy scour in the contracted channel during floods, cause the concave side of the cross-section to take a slope steeper than that of stability for the fine material of which the bed and banks of the river are composed, which results in the bank being undermined and in the mattress settling. This explanation is supported by the fact that on examining the mattress at a low stage of water it was found to be pulled apart near the water-line and in a few instances the f-inch transverse cables broken.

The damage done during the June rise was repaired in November, 1887, as well as the funds available for construction work would admit of, by tearing out or burning the bank mattress where it had been partly undermined, regrading the bank, and constructing 1,400 linear feet of mattress about 37 feet wide extending from a stage of 315 beyond the shore edge of the old mattress, also by covering the sloped bank with stone.

Repairs were also made from a to b, from e to d, from e to f, and from g to h, at which places the bank mattress had been slightly damaged.

These latter repairs could only be of a light character, on account of a lack of funds, and consisted in placing a shingling of brush fascines, covered with stone, behind the broken mattress, and regrading the bank with shovels and covering it with stone.

Statement giving the cost of the repairs made at Kaw Bend in November, 1887.

In March, 188, the bank mattress again began to settle between T and R and between O and W.

At the former locality this action was of only temporary duration, and the damage, which was slight, was repaired by the Kansas City, Saint Joseph and Council Bluffs Railroad Company having placed in the break a shingling of brush weighted with stone. These repairs were made by the railroad company, as their track appeared to be in danger of being washed away.

From O to W the revetment has been entirely washed away, and on June 29 the bank had cut back to the line O-N.

The following is a brief review of the history of the works in Kaw Bend: The protection of the bend was begun in 1879, in which year about 4,400 linear feet of light revetment of an experimental character was constructed near the middle of the bend. By October, 18-2, about 75 per cent. of this work had been destroyed by the river, but as late as July, 1883, the shape of the shore-line in the concave side of the bend was even and quite favorable for being protected by revetment. The appropriations up to 1882, however, had been made too irregularly, and were too small in amount to do the necessary work for protecting the whole bend.

In the fall of 1882 this part of the river was placed in my charge, and by June 30, 1883, about one and a half miles of bank was protected by revetment. These works

were situated near the middle of the bend, where at the time the greatest necessity for them appeared to exist.

Early in 1883 a project for completing the protection of the bend was prepared, but owing to the failure of Congress to make appropriations for rivers and harbors that year, this work could not be proceeded with, and the opportunity for holding the bend at the proper time was lost. The work was not resumed until the spring of 1885, and meanwhile, owing to the extensive erosions in the upper part of the bend, a part of the work constructed in 1882-'83 was flanked by the river and destroyed, and the shape of the bend entirely changed and made unfavorable for revetting.

By a further delay the conditions would probably have become still more unfavorable, and, besides, the interests at stake made it desirable to prevent the formation of a threatened cut-off by the river.

The experience gained by the failure of a part of the Kaw Bend revetment shows that this class of protection work can not be depended on to withstand very sharp impacts of the current or to be effective in bends of short radius, at least when accompanied with such other unfavorable conditions as in this case. It is quite improbable that such a state of things will exist, however, when the protection of bends is taken in hand while they have a comparatively regular outline and are of approximately even curvature. The shore-lines, channels, etc., of 1882 and 1885 at Kaw Bend are shown on Plate V.

East Bottoms repairs.-On account of the extensive erosion of the right bank at East Bottoms during the past few years the river has become so wide that its regimen is unstable, and during middle and low stages of water the channel is divided and constantly shifting.

In the latter part of August, 1887, a reef formed near the lower end of the revetment constructed the preceding June (see Plate UI, Fig. 2). A new channel opened around the head of this reef, and impinged on the revetted bank at an angle with its general direction of about 40 degrees. The upper end of the reef gradually cut away and the point of impact correspondingly shifted down-stream; simultaneously the reef widened, reducing the width of the chute and deepening the channel along the bank. This action lasted for about three weeks, and as a result the flexible mattress below the water-line settled, and the bank immediately behind it, from C to D (see Plate III), a distance of about 1,100 feet, was undermined.

As the stage of water decreased the channel above the reef shoaled, the reef was cut away by the current through the chute following the bank, and the undermining of the revetment ceased.

Soundings made September 7 showed the side slopes of the revetted bank below the water to be from 1 on 14 to 1 on . The sketch given on Plate III, Fig. 1, shows the change in shape of the cross-section in the vicinity of the break between July 2 and September 7. Upon examining the mattress at low water it appeared in good condition, with the exception of being torn near the water-line.

The repairs to this work, which were made in November, consisted in burning off the damaged bank mattress, regrading the bank, and constructing a mattress 1,137 feet long by 35 feet wide, extending from an elevation of mid-stage out over the inner edge of the old mattress.

A recent inspection of this work showed that it was in good condition. Strengthening the dike at East Bottoms.-The results accomplished by the dike at East Bottoms, constructed in the spring of 1887, are entirely satisfactory, the chute having been almost entirely silted up by deposits both above and below the dike to a stage of about 314 (K. C. gauge).

The area of accretions above the dike amounts to about 37 acres and below to 41

acres.