diate zone of varying width covered by "connecting mats," discontinuous in structure, but firmly joined to the upper work and to each other, and lapping well onto the subaqueous mat.

In construction especial attention was given to the internal strength of all the work, to secure junctions among the several parts, and with the old work above and below, and to anchoring to the bank both the subaqueous and shore revetments. As the defects in the present work in this locality have resulted mainly from a tendency of the upper revetment to slide down the bank, the latter point (anchorage) was especially considered, and with a view of increasing the range of this experiment, as also in the hope of greater security, anchor cables were run out from the old work below No. 4.

An experiment in drainage was contemplated, and directions given for the construction of a drain. The drain was begun at any early stage of the proceedings, but its completion was deferred until the river had risen several feet from the lowest stage. It was then hastily finished, but in such shape that it cannot fairly be credited with any good results which may follow, or charged with any bad ones.

No important changes were made in the methods of administration. The late date at which the work was authorized left no time for advertising, for which reason purchases of materials and supplies were made in open market, except the stone, which, not being required until near the close of operations, was obtained on requisition from the secretary of the committee on construction.

The 1,140 linear feet of bank in fault No. 4 were completely revetted at a total cost of $16.60 per foot. Several causes operated to increase this cost, among them the small amount of work to be done, the necessity of bringing plant from Plum Point and returning it; the serrated outline of the bank, requiring in the pockets nearly double the ordinary amount of shore work, and the fastenings to the old work above and below, which were tedious and expensive operations.

This revetment, for which $16.60 per foot of bank may be taken as a liberal estimate, needs only the addition of 25 feet width to the subaqueous mat (175 feet instead of 150 feet), to become what, in the light of present experience, I would recommend as a standard for localities of ordinary importance in this part of the river. The addition of 25 feet would mean an increase of 16 per cent. in the cost of subaqueous work, which is now 49 per cent. of the whole. It seems safe, therefore, to estimate such work at $18 per foot for the future.

The reballasting required 3,000 yards of rock, of which 280 yards were placed in the Memphis revetment, immediately below Wolf River, and the remainder on the Hopefield work, below fault No. 4. A classified statement of expenditures for this work is herewith.

It was thought that with No. 4 repaired no apprehension need be felt for the safety of this work during the succeeding high water. This prediction has been fully verified, since all the rest of the work is to-day in practically the same condition as it was a year ago. The lowest fault near the end of the work has enlarged slightly, but has not caved back, and there has been a loss of a few feet at the extreme end, due to caving working up from below.

The Memphis revetment is in perfect condition; but its completion to the end of the paved levee is of the greatest importance.

The changes in the harbor of Memphis have been marked and beneficial. The corner of the sand-bar, which extended out from Hopefield Point, and around which the river made a very abrupt turn, has entirely cut away, as intimated in last report. The river now follows a more gradual curve, striking the Memphis front less squarely. The engorged section is now at the foot of Beale street. Memphis Harbor, above the Elevator, is now comparatively slack water, and some filling is taking place. These changes are entirely in the direction contemplated by the Commission when the plan for this improvement was adopted four years ago.

Two small breaks are reported to have occurred in the Long Lake Levee in this district during the last flood. No funds have been available for an examination of them. The following financial statement and general balance-sheet show the present condition of appropriations for this district:

Appropriation for improving Mississippi River (act of July 5, 1884).

SECOND DISTRICT.

Balance on hand November 30, 1885, as per last report, including estimated liabilities.

$29,787 93

Dr.

Improving Mississippi River, Second District, Memphis Reach and Harbor.

CR.

Approximate value of plant belonging to the United States and used upon the improvement of the Mississippi River, Second District.

APPENDIX G.

REPORT OF CAPTAIN WILLIAM T. ROSSELL, CORPS OF ENGINEERS, UPON OPERATIONS IN THE THIRD DISTRICT.

UNITED STATES ENGINEER OFFICE,

Memphis, Tenn., July 26, 1886.

COLONEL: I have the honor to inclose herewith report on third district, Improving Mississippi River, for the fiscal year ending June 30, 1886. Very respectfully, your obedient servant,

Col. Q. A. GILLMORE,

WM. T. ROSSELL,
Captain of Engineers.

Corps of Engineers, President Mississippi River Commission.

REPORT.

During the past fiscal year the third district, Improving Mississippi River, was in charge of Capt. Clinton B. Sears, Corps of Engineers, until May 14, 1886. On this day he was relieved by Capt. William T. Rossell, Corps of Engineers, by virtue of S. O. No. 89, Par. 2, Headquarters of the Army A. G. O., Washington, April 16, 1886.

I.-LAKE PROVIDENCE REACH.

This reach is thus described. It extends from Carolina Landing, 517 miles below Cairo, to the foot of island No. 95, a distance of 35 miles.

For an account of the physical character of this reach the proposed improvement and work done previous to June 30, 1885, reference should be made to the reports of the Mississippi River Commisson for past years.

During the fiscal year ending June 30, 1886, all the work done, except care of public property, has been repairs to the Baleshed system of dikes and the building of an additional dike across the chute on range 66, the whole work consisting in closing the gaps left in Stack Island cross dike and in dike No. 11 the previous season; building a new dike across the chute on range 66, and extending the dike on range 69 across the chute to the bar on the Mississippi shore.

Dike No. 10, on range 66, was constructed of four rows of piles securely braced, with a heavy footmat sunk between the rows of piling. Dikes 11 and 13, on ranges 67 and 69, were of two rows of piles braced, but without footmat.

The repairs to the Stack Island cross-dike were made by re-enforcing the ends with three rows of piles and driving three rows across the gap in front of the dike, the whole well braced and a heavy footmat sunk across the gap.

These repairs were begun November 17, 1885, and finished January 11, 1886. The amount of the allotment was $13,500, and the amount expended $12,881.34. The small amount of work done was due entirely to lack of funds. The methods used were those heretofore described.

CONDITION OF WORKS.

Duncunsby system of dikes.-This system has nearly disappeared, only portions of cross dikes 6, 7, and 8 remaining. They have, however, served the purpose for which they were constructed, viz: reducing the volume of water through Skipwith chute, and thereby preventing its enlargement. The contraction has caused an increase of channel depth here.

Cottonwood system of dikes.-This system remained in good condition until April, 1836. Early in this month the longitudinal dike was broken by a raft or barge, and these breaks have since enlarged, and about 300 feet of cross dike No. 3 has washed away. This system has caused the bar below to increase both in length and height, concentrating the water and fixing the channel near Mayersville Island.

Mayersville system of dikes.-During the last high water a large amount of drift collected in front of the Mayersville cross-dike, and finally broke the piles at about the 20foot stage. There remain, however, large quantities of drift embedded in the sand in front of the dike and extending across the chute.

The Mayersville system of dikes has caused a general shoaling the entire length of the chute.