" caving " action, as it is called, generally goes on for several miles
tliroughout the bend, A B. The velocity is not immediately checked to
a considerable extent, the acceleration due to the fall contending with
the resistance of the bank. Each "pocket " excavated (shown by the
dotted lines) a^ords an additional point of attack to the water from
above; the load of sediment becoming accumulated and partially dis-
tributed as the different tillets, a, b, c, d, etc., come in contact with the
bank, or with the fillets next the bank, and from the cross and vertical
currents arising from the violent disturbance. Finally a jjoint will be
reached when the destructive energy is exhausted, the velocity reduced,
the angle of presentation of the bank less, and all the conditions favor-
able to deposition. The load will be thrown down here, forming a
shoal. The cross-section will become shallower, therefore wider, and
on the whole (the velocity being now at a minimum), enlarged consider-
ably. The slojie from C to D will, therefore, be steepened, and the
velocity gradually accelerated, until it again becomes normal, or a little
more, at D, where the same process will be repeated.
It is evident, then, that the tendency of the river, if unchecked,
would be to go on lengthening the bends. This would gradually give
a flatter slojDe to the whole stream, so flat that it would not be com-
patible with a velocity sufficient to carry the sediment to the sea. There
would, therefore, be a deposit in the upper jDart of the alluvial ijortion.
Ob STABLING ON THE MISSISSIPPI KIVER.
and a raising of the flood-plain ; and this process would go on indef-
initely, assisted by the lengthening of the stream due to the advance of
its delta into the sea.
A check is given to this development by two agencies.
First. â€” The eroding itself cannot jiroceed beyond a limited extent,
for it almost ceases when the bank attacked is at less than a certain angle
with the current. If water loaded to its normal capacity flow through
straight banks, it will dislodge no more. Neither will it at a gentle
angle. In short, it is found by observation that the caving jirocess does
not i^roceed to any considerable extent in ordinary soil and with the
ordinary high-water velocity when the radius of curvature of the bend
exceeds about two miles.
Second. â€” The river thus works out for itself a bend which would be
circular if the material were homogeneous, and the velocity continued
unchanged.' It has just been explained why it is not immediately dimin-
ished below the normal by impact against the bank â€” viz., on account of
the acceleration due to the fall from the shoal above. But even this
acceleration is presently overcome, the velocity is lessened, the curve
becomes flatter, erosion ceases, and the current leaves the bank which it
has so far hugged, with the direction given by the last reach, to cross
over to the other side. If this were not so, the river would speedily
return upon itself and form a complete circle. As it is, this tendency is
only partially overcome. The river does eventually almost return
upon itself until the neck/ 7 becomes exceedingly narrow. Sometimes
the overpour during high water across the narrowest part of the neck
forms a deep gorge on the lower side, which cuts back year by year
toward the upper side, after the ordinary manner of cataracts. During
some great flood, the velocity of the water, now precipitated across
the neck with a fall of perhaps 4 or 5 feet in a thousand, becomes
irresistible, and opens a passage for the whole river, thus forming what
is called a cut-ofi".
There is another cause which favors the formation of cut-offs. On
account of the continual erosion of the bank and bottom in concave
bends, the water becomes very deep, and a cross-section is formed so
favorable to the movement of water, that a comparatively small slope is
sufficient to give it the velocity requisite. This velocity is at length
partially exhausted by the resistance of the curved bank, until finally,
in the course of a long and deep bend, the river, even in flood-time, is
STARLING ON THE MISSISSIPPI KIVER. 89
aijproximated to the condition of a pool. Heuce, there is a tendency to
concentrate the fall into the reaches above and below the i^ool, so that
the slope is alternately greater, less and greater than the mean. Thus
the disjiosition to erosion at both the upper and lower sides of the neck
So, also, the formation of a quasi-pool in the bend creates a tendency
in the water to seek more direct routes. As the river pilots exjiress it,
"When a bend becomes very round, the water wants to leave it."
Now, almost all "points " have been formed by successive rudimentary
islands, which afterwards grow into perfect islands, become covered
with vegetation, and eventually are raised to nearly mean flood-mark.
These islands are divided from the main land by channels called "chutes,"
which are never entirely obliterated, but generally serve as highrwater
conduits. Frequently, in a bend which has grown very "round," the
river will seek such a short cut and make it the main channel, thereby
shortening its course several miles. See Plates VII and VIII.
It should be observed that these changes do not occur periodically,
or in obedience to any rational demand, so to speaik, for compensation.
They are local phenomena. Sometimes they happen unseasonably, in
advance of any undue lengthening of the river, or when it has already
been shortened by other cut-offs. In fact, it has been observed that
they frequently occur in cycles, several in succession, where a number
of points exist in close proximity, presenting favorable conditions for
cut-oflfs. See Plate IX.
Lest the river should become unduly short by these means, it pro-
ceeds to lengthen itself with great raj^idity by increased erosion of the
bends and building out of the points in the vicinity of the cut-off.
In these ways the disposition of the river to keep on increasing in
length is compensated, and the total mean distance from the head of the
alluvial basin to the beginning of the delta preserved nearly unaltered,
subject only to oscillations. The Mississippi has been observed with
accuracy for only a few years, but there is no evidence of any material
general elevation of the bed or lengthening of the course since atten-
tion has been directed to it.
This sketch of the regimen of the Mississippi implicitly indicates
both the nature of the evils which it is desirable to overcome or
ameliorate, and the means which lie at the disposal of the engineers for
such a purpose.
90 STARLING ON THE MISSISSIPPI RIVER.
The principal evils are :
First. â€” Obstruction of navigation by shoals;
Second. â€” Injury to alluvial lands by overflow ; and
Third. â€” Destruction of banks by erosion.
I. â€” The formation of shoals at the nodes or reversion points of bends
is assisted by the obliquity of the current to the general direction of the
river as it crosses from one bank to the other, thus affording a wider
cross-section and a greater frictional resistance.*
Besides these shoals, there are others, which make their appearances
princiiDally in reaches which are either straight or of small curva-
ture. These appear to be aseribable to the prolonged fall not now-
impeded by resistance from bands, resulting in excessive scour,
â€¢which is now directed against the bottom, and is followed sooner or
later by a checking of the velocity in the thread of the stream, and a de-
posit of silt in the bed below. Where the current and the erosion have
been greatest, there will be the greatest fill â€” viz., in the middle of the
stream. This will force the river to divide itself into, two branches, and
as the form of cross-section is unfavorable to the passage of the dis-
charge, there will be an enlargement of the cross-section. The current
being now deflected toward the banks, the enlargement will be at their
expense, and the river will assume the aiDpearance of a wide and shallow
stream, split in two by a princiijal shoal or series of shoals, while the
material derived from the erosion of the banks goes to form other sec-
ondary shoals below. The river is the more ready to shift its channel in
straight reaches from the lack of any centrifugal force to hold it against
Of the two channels, one is frequently used exclusively during high
â– water, the other at low and medium stages; the shorter arm being usu-
ally that preferred by the high-water current, according to a princijile
l^reviously stated. f The channel in use is marked by a rapid current;
the "dead end " being almost stagnant; the water level being approx-
imately that of the lower end thereof. In each case, the drsused chan-
* Jacquet on the Improvement of the Rhone, page 36. This memoir is pub'.ished with
three others : Janicki on the Improvement of N.ivigable Rivers ; Pdsqueau, A Project
for the Improvement of the Rlione ; aud Jacquet on Submerged Spurs ; by the
Engineer's Department, under the title, Improvement of Nou-tidal Rivera. Translated bj-
Colonel Merrill, Washington, 1881.
t AnU, page 89. A's3 noted by Jacquet, page 3G.
STARLING ON THE MISSISSIPPI RIVER. 91
nel deteriorates until the recurrence of the conditions which bring it
into play, when it soon scours itself out again.
Whichever class a shoal under consideration may belong to, the rem-
edy which at once suggests itself is the same â€” viz., a forced contraction
of the width of the cross-section. The result of this will be an imme-
diate rise of the water surface on the shoal. If the bed were stable, a
rapid would be thus produced. If it be erodable, the effect will be a
scour and a deepening of the channel to an extent commensurate with the
In a situation similar to that described above, contraction is most
easily effected by closing one of the two channels. In other situations,
artificial obstacles must be placed in the bed of the stream adjacent to
one or the other bank.
All contraction works are modifications or combinations of two sys-
tems, that of spur-dikes and that of training-walls. The former are
perpendicular or inclined at an abrupt angle to the course of the river,
and the latter are parallel or nearly so to that course. There are ad-
vantages and disadvantages attached to each of these systems, when ex-
clusively employed, which have often been set forth at length. Many
of the drawbacks may be avoided by a judicious combination of both
When circumstances admit, these structures are built of strong and
enduring material. On the Mississippi such materials are altogether
excluded on account of their expense and their great weight. The bot-
tom of that river is not only friable, it is extremely soft and treacherous,
and ponderous structures would soon sink out of sight, unless built on
foundations constructed with siDccial care and at immense cost.
Of course, a great part of the skill of an engineer lies in his ability to
make use of those materials which are nearest at hand, exist in the great-
est abundance, and are most easily worked with such labor as is avail-
able. These considerations have led to the adoption of pile-dikes con-
structed on the permeable system.
By permeable dikes are meant those which are expressly designed
not to exclude the water altogether, but to suffer it to pass through
them, opposing to that passage, however, a considerable resistance by
which a fill is induced. By combining a number of such dikes, system-
atically arranged, it is designed that an artificial bank shall at length be
produced, which will in time grow to the height of the natural, and
92 STARLING ON THE MISSISSIPPI RIVER.
whose other dimensions shall correspond substantially to the limits of
the system. This is more or less the object of all contraction works,
even those of stone.
The plan of permeable dikes presents many advantages on the Mis-
sissiiJjDi. It economizes material. It uses those articles which are
cheapest and most abundant. It is easy of construction and requires
little skilled labor. It is light, and remains in jDosition undisturbed,
even with the worst foundation. It is by far more effectual than any
other in inducing deposits. A standard objection to spur-dikes when
of massive construction, has al^vays been the slowness with which they
build i;p the spaces between them. For this reason, the German en-
gineers, the great experimenters with this system, were wont to recom-
mend that spur- dikes should be built, not at once to their full height,
but in successive layers, allowing suflScient time to elapse to allow the
alluvial deposits of different years to connect with each other.*
The iiermeable system allows continual access of fresh water heavily
loaded with sediment to the interior of the Avorks, thus protluciug de-
posits of enormous magnitude, thirty feet or more in depth in jalaces, in
a single year.
It must not be forgotten that the closing of a chute by a system of
dikes or any artificial narrowing from one side speedily involves en-
croachment on the opposite bank, unless the latter be protected in some
The protection of banks so unstable as those of the Mississippi is a
task which may well tax the best jDOwers of the engineer. Nevertheless,
it has been attempted with very tolerable success, the results varying
somewhat with the methods employed, the difficulties encountered and
the workmanship displayed. The greatest difficulties are a swift current,
a direct impact, a considerable depth and a very friable material. In
what is called "a very bad caving bank," a strong current strikes at an
abrupt angle a bank of nearly pure sand. The methods employed for
direct protection are, so far, only two â€” namely, continuous revetment
and submerged spurs. They will be described hereafter.
In the works constructed for the improvement of the Mississij^pi,
both spurs and longitudinal dikes are used, combined in systems. The
* Janicki, page 10. Scblicbting on the Improvement of Navigable, Non-tidal Rivers.
Translated by Lieutenant Frederick V. Abbot, Corps of Engineers, Willets Point, Battalion
Press, New York, 1885.
STARLING ON THE MISSISSIPPI RIVER. 93
genei'al principles which have been followed will be learned from the
annexed map of the Plum Point Reach.* This is one of the two great
"reaches" selected by the Mississippi River Commission for improve-
ment as containing the worst obstructions to navigation on the river, the
other being Lake Providence Reach,
The project for the improvement of the Plum Point Reach embraced
several distinct operations:
First â€” The closing of the chutes between Elmot Bar and the Ten-
nessee shore, by the main and cross dikes at and below Gold Dust.
Second. â€” The protection of the Arkansas shore from Fletcher's Land-
ing to Elmot.
Third. â€” The closing of the chutes between the Arkansas shore and
Osceola Bars and Bullerton Tow-head, and the i^rotection of the river
banks of the latter island.
Fourth. â€” The encroachment of the Tennessee shore upon the river-
l)ed by main and cross dikes below Plum Point.
Fifth. â€” The protection of the Arkansas shore below Bullerton Tow-
head against further caving, f
The history of these works is highly interesting and instructive, but
<5annot be given here. They were carried on under many discourage-
ments, the chief being the failure of the appropriations for several
successive years. The project has only yet baen partially carried out,
and last year, for the first time in the history of the work, the discharge
â€¢of the river, at its lowest stage, was confined to the regulated channel. J
Yet the last repoi'ts inform us of the permanence of the work and the
happy results obtained; the least depth of water reported since the
inception of the regulation being 8.5 feet, and that only for a few
In the case of the Providence Reach, the work of improvement was
complicated by a condition which often exists in the work of river
regulation; that is, the presence of a badly caving bank above the reach,
the erosion reaching further and further down every year, and threaten-
ing iierpetual change to the portion of river below, both from the
new set which was thus continually given to the current, and from the
material which was thus supplied for building new shoals. The
* Plate X.
t Report of Arthur J. Frith, Assistant Engineer, in Report of the Mississippi River Com-
mission for 1884, page 214.
t Captain Leach, in Report of Chief of Engineers for 1888, page 2196.
94 STARLING ON THE MISSISSIPPI RIVER.
advance of bars or shoals lias long been recognized* as a formidable
obstacle to regulation, and it is necessary to use some device to
prevent it, or it threatens the permanence of the whole system. In this
instance, the expedient adopted was the revetment of the caving bank in
Louisiana Bend, as it is called. Unfortunately, the work was left with-
out repairs for four years for the want of an appropriation, so most of it
has been lost. It will be rebuilt during the coming year. In other
respects, the work at Lake Providence Reach was very much of the
same character as at Plum Point, and similar results have been
Both longitudinal and cross dikes are built of several rows of
piles, connected with each other by stringers, braces and ties, as
shown by the annexed drawing (Plate XI). It may be stated, in this
connection, that the first constructions on all these works were on alto-
gether too economical and light a scale, and the materials too perish-
able. There has been a steady advance toward strength and durability
as experience has jirogressed. Cypress piles are used instead of cotton-
wood. There are four or five rows of piles where there were formerly
one or two. Drift-bolts have been discarded, as tending to split the
timbers, and wire fastenings substituted. These latter 'are of almost
universal use in all departments of river work. The wire is annealed
and galvanized, and is tightly twisted by a rack-stick. The piles are
driven with the assistance of a hydraulic jet.
The bottom of each dike is protected from scour by a "foot-mat "
made of brush and poles and weighted by stone in the same manner as
the mats for bank revetment. Experience has shown this precaution to
In many cases, the dikes are wattled with brush, to check the current
still more effectually. In the earlier stages of the work, inclined screens
of wire or brush were used for the same purpose, but I believe they
have been abandoned.
The vast ''systems" thus constructed at Plum Point and Lake
Providence, each covering many square miles, have built up enormous
tracts and have fulfilled the essential purposes or which they were
designed. But it is found that the interior parts of the system do not
jjrogress at the same rate as the outer joart. Minor channels and water-
* See Schlichting, page 136.
STARLING ON THE MISSISSIPPI KIVER. 95
-ways are preserved, wliich give no little trouble, and cause numerous
breaches in the dike, esiDeeially the longitudinal, in time of flood. A
system like that at Gold Dust (Plate X), connected with the bank at the
upper end and joined to an island like Elmot Bar, soon forms a basin
with raised margins, which lie along the head and the longitudinal dike,
from w-hich at moderate stages the water along the front is altogether
excluded, its only means of access being at the foot of the basin. The
level of the water in the basin at its head is then lower than that of the
river in its front by an amount ec[ual to the fall of the river from the
head of the system to the foot of the island. Hence, in time of high
water, results a tremendous overfall over the longitudinal dike, which
may cause any amount of destruction.*
It has been proposed by Major Ernst, one of the members of the
Mississippi Eiver Commission, who has had a great deal of experience
on this kind of work, to build the cross-dikes, not to their full
length at once, but by successive annual additions, as the area becomes
The contraction works, as a whole, are successful examples of what
may be done in that way; but there are arguments urged against con-
traction as a i^rinciple which are irrespective of the practicability or
success of the works themselves. These arguments are mainly two:
first, that the material scoured out in consequence of the contraction is
deposited below and forms other shoals ; and secondly, that the effect
of deepening a jiassage through a bar is invariably to lower the surface
of the pool above.
The former of these objections is not a serious one, nor is it neces-
sarily an objection at all. The quantity of earth eroded from a bar to
form a V-shaped navigable channel is not very great, nor does it follow
that it must be deposited in an inconvenient place. It may be deposited
in the pool below, which has an ample, nay, a superfluously great, cross-
section. Such is the process usually pursued by the stream in its nat-
ural state As the water falls, it cuts out for itself passages through
shoals which are generally sufficient for navigable purposes without
artificial aid. It must be remembered that the b?d is the work of a
great river (that is, the river at high water), and designed for its own
needs; and a small (that is, a low-water) river is compelled to use it.
* See this lucidly explained in Schlichting, page 164. It bas caused a great deal of trouble
on the Mississippi.
96 STARLING ON THE MISSISSIPPI RI7ER.
Hence, the liigh-water process is to deepen pools aud build bars. The
low-water jirocess is to scour bars and fill pools.* This process is
directly beneficial, as it tends to restore the slope which has been lost
in the pools.
The second objection is a grave one. It has been demonstrated that
such a consequence is a necessity on theoretical grounds, f and it is
abundantly borne out by experience. The effect, of course, of lowering
the water surface in the pool above, is to make still shallower water on
the shoal at the head of that pool, aud thus necessitate correction there.
This correction would involve a similar Consequence; and it appears at
first that this process must continue indefinitely.
In rivers of great slope, like the Ehone, in which the alluvial bed is
thin and the permanent strata easily exposed, it is found that the evil
above mentioned is very real and serious, and it appears fatal, in such
situations, to the system of contraction. On rivers of the type of the
Mississipi^i it loses much of its weight. In the first place, it is found
by experience that a lowering of the water surface three or four feet
(by a ciit-oflf, for instance, or a crevasse) is hardly perceptible 100
miles above.;}: This is the case even at high water, aud is attributed by
some engineers to " the small function of the slope which enters all dis-
charge formulas." | This reasoning does not appear sufficient, nor when
tested does it yield a satisfactory result ; as a difference of 5 feet in
100 miles should give, in a river like the Mississippi, by Kiitter's
formula, a difference of velocity of 0.3 to 0.4 foot i^er second. It would
api^ear rationally ascribable, in part at least, to the disturbances of the
flow produced by the resistance of the bends, by which the velocity is
brought to a partial check and takes a new start every few miles. Now
there are frequently very long stretches of the Mississippi (one, for
instance, above Greenville) where there are no shoals which can prove
dangerous to navigation for more than that distance. A lowering of the
water surface, say, 6 feet, then, at Leland Bar, just below Greenville,
*See on this a paper by Professor J. B Johusou, ia the Journal of the Association of
Engineering Societies for July, 1884, page 172.
t See Janicki, 13, 18-19; Jacqiiet, 38-39; Pasqueau, 08 et seq.
t Humphreys and Abbot, pa^e 397 (Eclitiou 187C) See also the interesting testimony of
Captain Leach and General Comstook in tbe Keijort of the Senate Committee on the Improve-
ment of the Mississippi River, in 1888, pages 6-7 aud 130-132.
Â§ Extract from Ueport of Board of Engineers in 1879, quoted iu Report of Mississippi
Biver Commission for 1880, page 13; aud see Captain Leach's testimony, as above.
STARLING ON THE MISSISSIPPI RIVEFJ. 97
would probably have no detrimental effect â€” in fact, would be imper-
ceptible at Helena, a distance of 175 miles. While it has been shown
that a certain lowering of the water-surface must necessarily result,