Majority — Rutledge J.
delivered the opinion of the Court.
Kansas brings this original suit against Missouri to have determined their common boundary from the mouth of the Kaw or Kansas River northwardly, over a distance of approximately 128 miles, along the channel of the Missouri River to its intersection with Kansas’ north boundary line.
At the time of Kansas’ admission to the Union, January 29, 1861, the western boundary of Missouri followed the
After the filing of the suit a master was appointed. Extensive hearings were held. Both documentary and oral evidence was presented. The master has filed his report, which makes findings and conclusions in favor of Missouri. Kansas says these are contrary to the law and to the weight of the evidence.
The land in dispute consists of about 2,000 acres. This now lies on the Missouri side of the river toward the lower end of Forbes Bend. Kansas claims this land was at one time soil accreted to the Kansas bank, which an avulsive change in the course of the main channel has put back on the Missouri side; or, in the alternative, that the tract
1 Cf. Missouri v. Kansas, 213 U. S. 78 ; Missouri v. Nebraska, 196 U. S. 23 .
2 Act of Admission of Kansas, 12 Stat. 126 ; Kansas Constitution of 1859, Charters and Constitutions of the United States, Part I, 629, 630.
3 The complaint alleged disputes over the line at points along the river between War Department Survey Stations 399 and 405, at other points in Atchison County, Kansas, and at points along the river between War Department Survey Stations 510 and 515 (Forbes Bend).
4 Attempts at settlement by negotiation had been authorized by Kansas before this proceeding was begun (Laws of Kansas, 1939, c. 355). Apparently they were unavailing, and this suit was instituted. After it was begun, however, the parties agreed to a settlement with respect to all areas but this one and incorporated it in this record. It will be made part of the decree.
I
The States are not in dispute about the applicable law. They agree that when changes take place by the slow and gradual process of accretion the boundary moves with the shifting in the main channel’s course.[5] Likewise, they agree that a sudden or avulsive change in that course does not move the boundary, but leaves it where the channel formerly had run.[6]
However, the parties are sharply at odds over the facts and the conclusions to be drawn from the evidence. In view of this and since we think the facts as presented by the evidence are conclusive of the controversy, it becomes necessary to sketch them and to refer to portions of the evidence in order to give an understanding of the issues and the basis for our conclusions.
Forbes Bend lies between Doniphan County, Kansas, and Holt County, Missouri. The disputed boundary, according to the master’s findings, extends along the main channel of the river as it now flows for a distance of about five miles bending southeasterly from Channel Mileage Station 515 to Station 510 (as measured and marked in
5 Jefferis v. East Omaha Land Co., 134 U. S. 178 ; St. Clair County v. Lovingston, 23 Wall. 46 ; Nebraska v. Iowa, 143 U. S. 359 .
6 Nebraska v. Iowa, 143 U. S. 359 ; Missouri v. Nebraska, 196 U. S. 23 ; Oklahoma v. Texas, 260 U. S. 606 .
The Missouri River is a vagrant, turbulent stream. Its name reflects this character. The Big Muddy is said to carry more silt than any other river except the Yellow River in China. It is constantly changing its course within the region between its bluffs, shifting from side to side as natural forces work upon its flow. Expert testimony is that a change of conditions in one bend produces changes as great, or nearly so, in the next bend below.
The river flows around a big bend, known as Wolf Creek Bend, just before it reaches the mouth of Wolf Creek. Here it runs almost due south. It is conceded by all that in 1900 the river flowed southeasterly in a single channel from the mouth of Wolf Creek, hugging the Kansas bluffs throughout the entire course of flow to Station 510. As it presently flows, the river makes a wide arc, first to the left or Missouri bank in a course almost due east or north of east, before it turns sharply to the south again at a point midway between the bluffs, and follows this southerly course until it strikes the old main channel at the Kansas bluffs above Station 510. This bend now is in the form of a bow, with the river proper forming the bow and the old channel along the Kansas bank its string. Roughly,
However, as will appear, the channel’s present location results from more complex changes than merely a movement of the river north and east over the distance lying between these two channels. According to the greatly preponderant though not undisputed evidence, there was a division of channels in Forbes Bend from about 1914 or 1917 to 1927 or 1928. During this time the more westerly or Kansas channel lay slightly west of where the present channel runs. The Missouri or easterly channel lay on the other side of the area in dispute, which then formed part of a bar or island. At one time, probably about 1922 or 1923, during the period of greatest erosion of the Missouri bank, this channel came within half a mile or less of the Burlington tracks. The Missouri channel, with the river above it, then followed a course almost due east or slightly north of east from below the mouth of Wolf Creek to the point of its closest approach to the railroad. Then it swung sharply to the south and in a curving line came back to join the original channel near the Kansas bluffs above Station 510.
From the recital thus far it is clear that in 1900 the land which then lay where the disputed tract now lies was Missouri land. This is undisputed. Likewise, the tract now is attached to Missouri on the easterly bank of the river. This is because the Missouri channel dried up during some five to eight years beginning around 1927 or earlier. But, before that process began, for many years the land in question lay between the two channels. And it is from conflicting views concerning whether, how and when these major changes took place the parties derive their respective claims to sovereignty over this soil.
7 Since Kansas claims avulsive change both in 1917 and in 1927, and that the accretion began about 1900 or shortly thereafter, her claim necessarily implies that the period of accretion extended either from 1900 to 1917 or from 1900 to 1927.
8 At that time, according to this claim, the main channel of the river flowed through the so-called Missouri channel to the north and east, but was suddenly changed by the ice jam back from that channel into a chute on the Kansas side. This chute previously had cut across the allegedly accreted land a little to the west of where the present channel now lies. The complaint alleges that the ice jam occurred “on or about February 1918.” The scanty testimony in the record if completely accepted would establish the ice jam in 1917 rather than in 1918; cf. note 24 infra.
9 The complaint alleges that the ice jam occurred “during the year 1927.” The witnesses who testify to the jam at this time date it variously in 1927,1928 and 1929; cf. note 25 infra.
Missouri meets all of Kansas’ claims with denial on the facts. She says first that the land in question has been at no time accreted soil of Kansas. On the contrary, she claims that the disputed area formed as an island in the river bed beginning about 1910 or 1912, and from then until 1927 or 1928 there was a divided flow around this island, a Missouri channel running north and east of it with a Kansas channel to the south and west. She insists that the Kansas channel always remained the main channel of the stream and only a minor one reached proximity to the railroad tracks. Accordingly, she says the island formed as Missouri land and always remained Missouri territory. Missouri thus opposes her view of island formation, both to Kansas’ view of that process and to her claim of accretion and avulsion.
However, Missouri adds a further argument even if the Kansas theory of accretion is conceded. According to this, the effect of the accretion to the Kansas bank is counteracted by the fact that at no time was there an avulsive change, whether in 1917 or in 1927. On the contrary the river moved back gradually as it came. In this view the accretive process working against Missouri ended in 1923 or 1924, when the Missouri channel reached greatest proximity to the railroad tracks. Beginning in those years and continuing gradually until 1933 or 1934, the river moved slowly back to a point beyond the location of the present channel. Thus purporting to follow the accretion theory in both directions, Missouri claims the land in question.
It may be noted that crucial to Kansas’ case, whether on her theory of accretion and avulsion or on that of island formation, is the need for showing that the main channel followed the course of the Missouri channel.
II
Perhaps the most important documentary item is a map of the Forbes Bend region compiled from the field in 1923 by the United States Engineer Office at Kansas City, designated as complainant’s Exhibit 46 in this record. Another map of considerable assistance, with information penciled on it by two witnesses who testified at the hearings, is complainant’s Exhibit 47. This purports to show in less detail than Exhibit 46 conditions in the Bend in 1926. The witnesses’ penciled additions, placing channels and other landmarks, with their testimony, give considerable information about conditions in the Bend from 1921 or 1922 on to 1926 and later. Assistance also is derived from complainant’s Exhibit 56. This is an aerial survey photograph of the Forbes Bend region made in 1941, showing conditions when this suit was begun. Reference to these exhibits will be made as the testimony of some of the witnesses is referred to.
It is clearly established that sometime around 1900, fixed by some older witnesses variously as beginning earlier and by others later, the river began a northerly and eastward movement, cutting away the Missouri bank and filling in on the Kansas bank.[10] Neighborhood testimony attributes the beginning of this movement to some change
10 C. McWilliams (1892); P. Dyer (1898); C. Hudgins (1900); J. H. Simpson (1904); J. E. Simpson (1905).
The soil composition of the north bank is a common formation in the Missouri River valley. Underneath the surface soil is sand or quicksand. This is covered by a layer of gumbo soil. Testimony in the record discloses there is no great erosion when the water is very high or very low. But when it is at an intermediate stage the water comes in contact with the underlying sand, washes it out, and the topsoil falls into the river in great chunks, often twenty feet long by ten feet wide. As the current was forced from the Kansas rock to the Missouri sand, it undercut more and more of the Missouri soil. Evidence in the record also shows that between 1900 and 1920, or a little later, from 4,000 to 5,000 acres of Missouri soil was washed into the river by this process. On this stood houses, barns, a school building known as the Baker schoolhouse, and other structures, which either went into the river as the soil was undermined or were moved to prevent their falling in. The Baker schoolhouse, which in 1900 was a mile or more northeast from the river bank, was moved about 1915 to prevent its going into- the river.[12]
11 Cf. testimony of L. F. Stalcup.
12 Witnesses vary as to the exact time from 1910-11 to 1916 (J. H. Peret: 1910-12; Mrs. S. Jenkins: 1910-12; E. E. Simpson: 1913; C. McWilliams: 1912-13; C. Harper: 1916; C. Hudgins: 1915; B. Hudgins: 1916; E. McCoy: 1915). But most of them put this event in 1912 or later, and the most reliable testimony, by those who moved the building (C. Hudgins and B. Hudgins), places it in 1915 or 1916.
The clear weight of the evidence is that there was only a single channel of the river until about 1912 or 1914. Witnesses for both Kansas and Missouri substantiate this.[14] The evidence also clearly establishes that there was a divided flow from 1917 or earlier tO' 1927. This too is substantiated by both Kansas and Missouri witnesses.[15] The evidence, however, is conflicting concerning when the division first took place and whether, while it remained, the Kansas channel or the Missouri channel was the main one.
Witnesses for Missouri, and some for Kansas, testify that the division occurred before 1917 and that the two channels remained substantially equal or the Kansas channel was the larger in the volume of water carried between the time of the division and sometime between 1922 and 1927 or 1928.[16] The Missouri witnesses fairly uniformly agree that the flow in the two channels was at
13 C. McWilliams (1924-25); E. McCoy (1922); A. H. Murray (1922-23); Ralph Dyer (1923-24); W. Metcalf (1917); cf.E.A.Cole (1923,1928).
14 Varying dates are given for the time at which a divided flow was first noted. Kansas witnesses: I. Muse: 1900; L. F. Stalcup: 1910-1911; J. E. Simpson: 1912-13, 1917; 0. McKay: 1917; R. E. Simpson: 1918; C. Baskins: 1917; C. W. Ryan: 1917 or 1919; J. McKay:, 1920. Missouri witnesses: D. Barbour: 1903; B. Hudginá: 1914;' C. Harrison: 1914; Ralph Dyer: 1913-14; C. Harper: 1915; A. H. Murray: 1915; H. H. Hall: 1916; W. Metcalf: 1916; J. Fitzgerald: 1917; Raymond Dyer: 1917-1918.
15 See note 14 supra. A few Kansas witnesses maintain there was only one channel through this period.
16 See note 14 supra. Kansas’ witnesses testified variously that there always had been a channel on the Kansas side, that it was swifter than the Missouri channel (J. H. Simpson), that the Kansas channel *223 was the “main river” (Mrs. J. Coufal), that the Kansas channel was much the larger in 1918 (R. E. Simpson), that most of the water was on the Kansas side in 1920 (P. Bottiger); cf. C. B. Catón, that the river was just about evenly divided in 1917 (C. Baskins). Missouri witnesses said that there was always a substantial flow in the Kansas channel and that it was about as large as or larger than the Missouri channel (e. g., Ralph Dyer, B. Hudgins, C. Dinwiddie, J. Fitzgerald, C. Harper, Raymond Dyer). They placed boats in the Kansas rather than the Missouri channel (E. McCoy); in 1918 (W. L. Moore); 1916 to 1929 (H. H. Hall); and in 1927 (C. Hudgins).
17 See note 16 supra.
18 W. Metcalf: 1917; W.L. Moore: 1918; C. Dinwiddie: 1920-22; C. Harrison: 1921; J. Fitzgerald: 1923-1925; cf. P. Bottiger: 1920.
19 J. Fitzgerald, E. Wales, J. H. Peret (Kansas witness); cf. E. McCoy, C. Harper.
20 E. g., J. H. Gray: 1928 et seq.; A. F. Hays: 1926 et seq.; J. B. Gray: 1927-30; G. Atkinson: 1929-after 1934; J. H. Peret: 1929-33; C. Coufal: 1929-33; C. W. Ryan: 1928-31; some Kansas witnesses claim the drying up of the Missouri channel was a sudden concomitant of an ice jam in 1929 but add that the Missouri channel contained water until 1933 or 1934 (e. g., C. Coufal, E. A. Cole), or 1935 or 1936 when it dried up as a result of government diking and revetment work upstream (e. g., J. Coufal).
On the other hand, Kansas witnesses are not in accord among themselves as to what occurred in the Bend between 1912 and 1928. Some of them say there was a divided flow.[21] Others deny this but qualify their denials by asserting that, although the main channel of the river ran over into Missouri close to the Burlington tracks until 1927, there were chutes on the island and particularly there was one chute running from about the mouth of Mill Creek Ditch as it was in 1917 (directly across northerly from the northern end of the bar or island) due south to the Kansas bluffs at about Station 510.[22] However, they maintain generally that this was a small chute, or smaller than the Missouri channel, at any rate up to 1922 or 1923. A few say it was small until 1927 when the alleged ice jam occurred.[23]
Some witnesses for Kansas maintain that there were big ice jams in 1917[24] and in 1927.[25] Different witnesses testify to the two alleged jams. Those who say one occurred in 1917 assert that it threw the main flow back from the Missouri channel into the Kansas channel. Likewise some of those who say there was a jam in 1927
21 See note 14 supra.
22 E. g., A. F. Hays, C. Coufal, K Brownlee, K. Robinson.
23 E. A. Cole, J. Coufal. Cole is a Kansas claimant to ownership of part of the disputed land. Coufal once worked for him.
24 C. Baskins: 1917; J. E. Simpson: 1917; P. Dyer: 1917; C. Dyer.
25 E. A. Cole: 1929; C. Coufal: 1929; J. Coufal: 1929; Mrs. J. Coufal: 1929; E. L. Rockwell: 1927; I. Overstreet: 1927; H. W. Linville: 1927; P.Dyer: 1927 or 1928. Mrs. Coufal, however, testified the “main river” was on the Kansas side of the island at that time. In this respect her testimony flatly contradicts that of her husband and Cole.
When we turn to complainant’s Exhibit 46 we find very substantial support for Missouri’s view that during the controverted period the flow of the river was divided and that the Kansas channel equalled or exceeded the Missouri channel in the flow or volume of water carried. This map, compiled by the Corps of Engineers of the United States Army, who had charge of the river’s development, shows conditions in 1923. Two channels appear, with a large sand bar or island between them. The map places the Missouri channel within less than a half mile of the Burlington tracks. It shows a width of about 1,250 feet at the narrowest point. Soundings, read from the
26 D. Baskins, J. Kotsch, R. E. Simpson, J. H. Simpson, W. Prusman, G. Atkinson.
27 1. Muse, L. F. Stalcup; cf. A. P. Staver.
28 C. Hudgins, V. Harrison, C. Dinwiddie, R. L. Greene, W. Metcalf, E. Wales; cf. E. McCoy, H. H. Hall, D. Barbour.
29 Cf. note 20 supra; J. Kotsch, W. Prusman, C. McWilliams.
Exhibit 46 furnishes the most reliable evidence in the record of conditions in Forbes Bend at a given time. If only this exhibit and the facts it discloses were considered, clearly it could not be ruled that the main navigable channel of the river was the Missouri channel. For purposes of navigation, the Kansas channel was much the shorter and more direct and, from the soundings as well as the shorter flow, it apparently carried at least an equal volume of water.
Exhibit 47, which is described as a revision from airplane photographs, shows in general a somewhat similar though less detailed picture for 1926. However, two witnesses for Kansas, Kenneth Robinson and Joseph H. Gray, who hunted in the region from 1920 to 1927 or 1928, testified concerning this exhibit and marked on it in penciled lines their recollections of the channels’ respective courses in 1922. Their testimony gives perhaps the strongest support to Kansas’ case that the main channel
III
The evidence need not be stated in further detail. In our opinion it fully supports the master’s ultimate findings and conclusions. It was his view, first, that there was no avulsive change, whether in 1917, 1927, or at any time. He found there was some evidence of an ice jam in 1917 and more to substantiate such a claim for 1927. But he also found, and the evidence, though not undisputed, fully substantiates his conclusion, that neither of these jams was sufficient to cause a sudden change in the river’s course.
There is very considerable doubt, on the record as a whole, whether the alleged jam in 1917 occurred at all.
30 They agreed that the Missouri channel flowed around the island not far from the Burlington tracks, turning south at that point and flowing against the Kansas bluffs at Station 510. They also agreed that the Kansas channel was a chute. But they differed concerning its direction and location. Robinson placed it as running almost due south across the center of the island in a straight course. Gray placed the Kansas chute more to the west and with a curving course. Both testified that the Missouri channel was the main channel at that time. The inconsistency between Exhibit 46 and the testimony and drawings of Robinson and Gray may be accounted for in part, though not altogether, by the fact they were in the Bend for hunting and fishing purposes, chiefly in the fall, whereas Exhibit 46 was made from surveys in June and July. The difference in time, however, is hardly enough to account for the difference either in width or depth of the Kansas channel as shown by the exhibit and by their testimony.
We agree with the master that the evidence to show a more unusual piling up of ice at the head of the island in 1927 is somewhat stronger.[32] But we also agree with him that the evidence as a whole is clearly preponderating that this did not cause an avulsive change. As has been stated, there is some evidence that the alleged 1927 jam caused the main channel of the river to shift then and suddenly from the Missouri channel to the Kansas channel as the latter flowed from 1927 or 1928 until the Government’s revetment work on the river forced the channel to its present location after 1935. But this is not enough to sustain Kansas’ case.
Both by virtue of her position as complainant and on the facts, Kansas has the burden of proof in this case. Cf. Oklahoma v. Texas, 260 U. S. 606 . The disputed location was in Missouri in 1900. It lies on the Missouri side now and has done so, by practically all the evidence, since at least 1927 or 1928. These facts put upon Kansas the burden of showing that in the meantime the land lay on the Kansas side of the main channel by virtue of natural changes which were effective to change the jurisdiction. Kansas has shown beyond doubt that one branch of the river eroded to a point or points north and east of this land, probably as early as 1920, possibly earlier. But beyond this fact, whether on her theory of accretion and avulsion or on that of island formation, the weight of the evidence is against her view of what occurred.
31 Cf. notes 27-29 supra.
32 Cf. note 25 supra.
Kansas’ evidence concerning the division of flow and formation of the island, together with that concerning
33 Missouri v. Kentucky, 11 Wall. 395 ; Davis v. Anderson-Tully Co., 252 F. 681 (C. C. A.); Commissioners v. United States, 270 F. 110 (C. C.A.).
34 Missouri apparently urges that even if the land formed as an island on the Kansas side, the process by which the main channel shifted from the eastern to the western channel and the former gradually filled with alluvial deposits thus connecting the island to the Missouri shore, entitles it to sovereignty over the disputed lands.
Kansas’ stronger case upon the proof is on the theory of island formation. On this, as under that of accretion and avulsion, it was necessary for her to show that the Missouri channel was the river’s main channel and thus the island, which is now part of the disputed land, was formed on her side of the river’s thread. On this crucial issue Kansas’ case is stronger perhaps than in any other respect. She presented substantial evidence to show that while the river was divided or during some part of that period, more especially from 1921 or 1922 to 1927, the Missouri channel was the main one, both in volume of water carried and, less clearly, in availability for navigation. There is, however, at least an equal weight of evidence, given both by Missouri witnesses and by some for Kansas, that the Kansas channel remained the main navigable channel throughout the period of division.
The evidence on this controlling issue unfortunately is not as free from conflict or doubt as we might wish. But it cannot be said, when account is taken of all the evidence, both oral and documentary, that a preponderance sustains Kansas’ view that the main channel ever changed to the Missouri side. Kansas’ burden required preponderant proof. She has not made it.
As the case has been made, both the master and this Court have had to rely upon the inadequate and inconclusive documentary evidence and the conflicting and often vague recollections of neighborhood witnesses. The sum does not add up to the weight of proof Kansas was required to establish in order to prevail. The master saw and heard the witnesses. His conclusions in all respects were in favor of Missouri. We find no basis in the record for any conclusion that he performed his task with other than fair, disinterested, painstaking effort and attitude.
It follows the land in dispute must be awarded to Missouri and the boundary will be fixed as the master has recommended in his report. A decree will be entered accordingly. [See post, p. 654.]
Cases cited
- Oklahoma v. Texas 260 U. S. 606 p. 228