Trustee of the Internal Improvement Fund of the State of Florida et al., Appellants, v. Walter F. Nowak and Jean F. Nowak, et al., Appellees, 401 F.2d 708 (5th Cir. 1968)Annotate this Case
Clyde O. Martz, Asst. Atty. Gen., J. Edward Williams, Acting Asst. Atty. Gen., Roger P. Marquis, Robert M. Perry, Attys., Dept. of Justice, Washington, D. C., Francis G. Rearick, Orlando, Fla., J. Kenneth Ballinger, Asst. Atty. Gen., Earl Faircloth, Atty. Gen., State of Florida, Tallahassee, Fla., for appellants.
Gene Essner, James E. Tribble, Blackwell, Walker & Gray, Miami, Fla., for appellees.
Before THORNBERRY and SIMPSON, Circuit Judges, and SUTTLE, District Judge.
THORNBERRY, Circuit Judge:
In 1963, the United States instituted proceedings to condemn land in Brevard County, Florida for use in the construction of NASA Launching Sites. At the time of filing the declaration of condemnation, the Government was of the opinion that the tract in controversy, i. e., Tract 3837-A known as Billy Joe Point and having approximately 98.58 acres, was owned by either the State of Florida or Appellees, Walter and Jean Nowak. Accordingly, it paid $27,000 into the registry of court as fair compensation and awaited determination of the ownership question. After the district court had entered findings of fact and conclusions of law to the effect that Tract 3837-A was owned by the Nowaks, the Government reached the conclusion that it was owned by neither the State nor the Nowaks. On an unsuccessful motion for new trial and on this appeal, it contends that the United States has always owned the land in question. The State, represented by the Trustee of the Internal Improvement Fund, contends that it should receive the compensation because it is entitled to a patent of the land from the United States while the Nowaks urge affirmance of the judgment below. We affirm.1
In 1859, the Surveyor General of the United States retained a man named Harris to survey a number of townships in Brevard County. His surveys were to aid the Secretary of the Interior in the preparation of plats of land to be patented to the State of Florida under the Swamp Lands Act of 1850, 9 Stat. 915, 43 U.S.C. §§ 982, 983, the purpose of which was to enable the conveyance to the states of all swamp and overflowed lands made unfit thereby for cultivation so that they could be reclaimed by the construction of levees and drains. Rogers Locomotive Machine Works v. American Emigrant Co., 1896, 164 U.S. 559, 17 S. Ct. 188, 41 L. Ed. 552. According to the testimony of C. C. Jobe, a civil engineer and land surveyor, Harris surveyed some 20,000 acres in the general area designated as Township 22, Range 37. On the basis of his field notes, the Surveyor General prepared and approved a plat pursuant to which a patent covering more than 4,000 acres was issued to the State of Florida in 1883. The Nowaks are subsequent patentees. The close question raised by this case is whether the 1883 patent based on the 1860 plat conveyed to the State the 98-acre tract known as Billy Joe Point. If so, the Nowaks own it and are entitled to compensation; if not, it was never conveyed and no compensation is due.
Tract 3837-A lies in the northwestern part of a large island and projects into the Banana River. The following outline represents the northwestern part of the island as it appears on the 1860 plat:
NOTE: OPINION CONTAINING TABLE OR OTHER DATA THAT IS NOT VIEWABLE As the dotted line indicates, Tract 3837-A does not appear on the plat. Immediately south of the tract are two lots designated as Government Lots 5 and 6. From the plat it would appear that all of Government Lots 5 and 6 border on the Banana River though, in fact, only part of Lot 5 is on the river. For the most part, the two lots border on the southern edge of Tract 3837-A. The following drawing represents the actual situation:
NOTE: OPINION CONTAINING TABLE OR OTHER DATA THAT IS NOT VIEWABLE
The controversy arises because Tract 3837-A does not appear on the 1860 plat. The Nowaks argue that this makes no difference because the surveyor and those who prepared the plat intended to "meander" the river or follow the shoreline. As will be discussed hereafter, there is case law for the proposition that when a surveyor has attempted to meander a body of water, that is, make the boundary conform to the sinuosities of the shore or bank, the mean high water mark of that body of water becomes the boundary rather than the exact location of the line drawn on the basis of the surveyor's field notes. Thus, say the Nowaks, the Banana River was intended to be and is the northern boundary on the plat so that any land to the south, including Tract 3837-A, was conveyed by the patent. Appellants, on the other hand, contend that Tract 3837-A was not surveyed by Harris and was not shown on the plat and that land not identified by Government survey could not be conveyed under the Swamp Lands Act. To come to grips with these opposing arguments, we find it necessary to set out in full the findings of fact entered by the trial judge:
FINDINGS OF FACT
1. The land designated herein as Tract 3837-A was in existence at the time of the Harris survey in 1859. The only testimony on this point was that of C. C. Jobe, a civil engineer and land surveyor who testified in behalf of the defendant Nowak. Jobe surveyed and platted the subject tract as well as Government Lots 5 and 6 on April 22, 1960. This plat, introduced as Nowak's Exhibit No. 6 and admitted without objection, shows the northern boundary of Government Lots 5 and 6 which constitute the southern boundary of Tract 3837-A as a straight line designated as a meander line on the plat. On the original plat prepared on the basis of Harris' field notes and approved by the Surveyor General Francis L. Dancy, a copy of which has been admitted herein as Nowak's Exhibit No. 3, that same meander line appeared as a relatively straight line bordering on the Banana River. The land south of the line is indicated as being "Savannah". Other places on the plat are designated as marsh. There is no such designation north of this meander line. To anyone reading such plat it would appear only that there is a mass of water known as the Banana River directly north and adjacent to this meander line.
Jobe testified that Tract 3837-A presently exists as a low lying tract of land devoid of trees and dotted with small ponds of water. The Coast and Geodetic Survey of 1883, introduced without objection as Nowak's Exhibit No. 5, clearly shows that Tract 3837-A was in existence in approximately its present form in 1883. The Court notes that there are streams shown on Government Lots 5 and 11 on the plat of Jobe's survey and also appearing on the 1883 survey but not shown in any form on the plat of the Harris survey. The Trustees contend, in support of their position that Tract 3837-A appeared after 1860, that the river had receded between the 1860 and 1883 survey. The presence of these creeks in the 1883 survey and Jobe's survey lead to an opposite conclusion, unless it be further assumed that Harris by design or error omitted those creeks. In addition, Jobe testified that on the basis of the present condition of the land and his various soil samples it was his expert opinion that the subject tract in question had not been permanently under water in 1859 or at any time thereafter. The Trustees introduced no testimony to the contrary.
2. As to why this parcel of land was omitted from the survey the Court can do little more than speculate. Some light is shed on this question by the Contract of Survey and various correspondence by the Surveyor General to Harris which were introduced into evidence by the Trustee and admitted over objection. In a letter dated March 18, 1859, from the Surveyor General, Dancy, to Harris, it is specifically stated that in regard to the land to be surveyed:
"You will make an examination of it and should it be found that there is not sufficient good land or land fit for cultivation and not subject to overflow to pay the expense of survey — you will at once abandon it, as it is not proposed by your contract to survey any lands that are not fit for cultivation or that would not find ready sale when brought into market."
In view of both the above quoted directions and the present state of the land today as testified to by witness Jobe, that is, that it appears not to be suitable for cultivation at the present time and, except for condemnation purposes, there is a strong question as to even its saleability, it appears likely that the land was omitted intentionally from the Harris survey as land being not fit for cultivation. Error on the part of Harris, if there was an error, was his failure to designate the nature of the land north of Government Lots 5 and 6, that is, whether that land, was, in fact, impassable marsh, land subject to overflow, or land temporarily or permanently covered by water. It further appears likely that on the basis of the field notes made by Harris showing that the meander posts at the corners of Government Lots 5 and 6 were on the "right bank of Banana River", persons preparing the plat drew the meander line as though it bordered upon the Banana River without considering whether or not the land north of such meander line was submerged land covered by the Banana River or was merely land omitted from the survey because of lack of saleability or being unfit for cultivation. (Emphasis added)
* * * * * *
It must be remembered that the Government did not advance the argument it now makes until these findings had been entered. Having studied them, it concluded that since Tract 3837-A was intentionally omitted from the survey it could not have been patented to the State and ultimately to the Nowaks. In other words, it concluded that the court's findings pointed to a different result from the one reached. In attempting to untangle this problem, we begin with this important question: What did Harris intend to do when he made the survey? From a careful reading of the district judge's findings of fact and conclusions of law, we gather that he did not believe Harris' intentions to be important because the plat conclusively showed the boundary of the area in question to be a meander line along the Banana River. To understand the plat, we think some conclusion must be reached as to what Harris was doing.
One possible explanation of why he might have drawn a boundary line across the top of Government Lots 5 and 6, omitting any reference to the subject tract, is that the tract was submerged when he made the survey so that the river came about 900 feet south of where it now is. However, C. C. Jobe testified and the court found that the tract was not under water in 1859. There was no evidence to the contrary. Having eliminated this possibility, we turn to the lower court's speculation that Harris ignored the tract because of the Surveyor General's instruction that he not survey any land unfit for cultivation. We will not label this a fact finding because the trial judge admitted he was speculating and did not approach the case with the idea that what Harris intended to do was decisive. Our own view is that the Surveyor General's letter of instructions would not have caused Harris to survey the general area — Township 22, Range 37 — yet intentionally omit parcels of land here and there because they were unfit for cultivation. In the first place, it would have made no sense for the Surveyor General to direct him to disregard any and all land unfit for cultivation because the purpose of the statute pursuant to which the survey was being undertaken was to enable the conveyance of all swamp and overflowed lands made unfit thereby for cultivation so that they could be reclaimed. Secondly, the plain words of the letter do not admit of the interpretation urged by appellants:
Before proceeding to section any of the townships embraced in your contract, you will make an examination of it and should it be found that there is not sufficient good land or land fit for cultivation and not subject to overflow to pay the expenses of surveying — you will at once abandon it, as it is not proposed by your contract to survey any lands that are not fit for cultivation or that would not find ready sale when brought into market. (Emphasis added).
Harris was to make a preliminary judgment as to whether a particular township contained enough good land to pay surveying expenses, and if he found enough good land, as he evidently did in the case of Township 22, he was to proceed to survey all of it. In light of the purpose of the statute and the language of the letter, this is the most reasonable interpretation. A final point in this connection is that the soil condition of Tract 3837-A is the same as that of the rest of the area so that Harris would have had no reason to omit that particular parcel because it was unfit for cultivation.2
Thus far we have said that we do not believe Harris had any reason for "intentionally omitting" Tract 3837-A from his survey. Stated differently, if, as he made his survey, his intention was to follow the sinuosities of the riverbank, he would have had no reason to deviate from his pattern when he reached 3837-A by drawing a boundary line across the top of Government Lots 5 and 6 to indicate that no land to the north would be conveyed. Having said this, we must face up to the fact that neither the field notes nor the plat suggest the existence of the subject tract. Appellants would argue that regardless of the reason, Harris obviously intended to establish a northern boundary for Lots 5 and 6 rather than follow the river around 3837-A. This argument is compelling so long as attention is focused on that part of the plat where a relatively straight line borders Lots 5 and 6 on the north, but our conception changes as we analyze the entire plat. A section from the field notes cited by appellees is instructive in this regard. In reading the following excerpt from the notes, it is helpful to refer to the first drawing set out in this opinion (the large one), keeping in mind that Section 33 is in the lower right-hand corner of the drawing and that Sections 27, 22, 15, 16, 9, 4, and 5 move up the right-hand side and around to the left. Section 6, the last one mentioned, is in the area where 3837-A is located:
Meanders of the Right Bank of Banana River, the East & Part of the North shore of Merritt's Island Begin at Meander Corner on the South boundary 53.00 chs. East of corner & on the Right Bank of said River & Run then up stream with the Meanders of the Right bank of the River in Sec. 33 as follows: (Emphasis added)
* * *
Thence in Sec. 27
* * *
Thence in Sec. 22
* * *
Thence in Sec. 15
* * *
Thence in Sec. 16
* * *
Thence in Sec. 9
* * *
Thence in Sec. 4
* * *
Thence in Sec. 5
* * *
Thence in Sec. 6
Standing on the right bank of the river in Section 33, Harris indicated that he would follow the meanders of the river northward and westward around the island. Of course, as he proceeded he measured off straight lines or traverses, but his intention was to follow the meanders of the river and for those who prepared the plat to show the river as the boundary. The notes do not reflect a different intention with respect to Section 6, and, as we have said, there was no reason for him to have a different intention.3
Having reached a conclusion about the surveyor's intention, we are prepared to deal with the next question: Why did he not measure off some lines or traverses which would reflect the existence of a knob of land projecting from Lots 5 and 6 into the river? The answer lies in Jobe's testimony, the gist of which is that Harris would not have gone to the trouble of following the river around Tract 3837-A because of the vast amount of land to be surveyed, the price for which the survey was being made, and the difficulty of making it, particularly the difficulty encountered on the soft land near the water:
I would be glad to tell you my basis for the answer of the question as I understood it, was the fact that the man was paid six dollars a mile for the survey, would affect the accuracy of his survey — my answer is it definitely would because surveying is divided into degrees of accuracy.
For instance, we have now the first order of work, the second order, and most land surveys are performed under a third order, work which is a horizontal enclosure, unadjusted field enclosure of 5,000 feet, where you get into higher accuracy.
The first order of accuracy is one to 50,000. The accuracy of the survey to control space problems is one to 52,000. Obviously, the instrumentation and time it takes to do this work is certainly the measure of the accuracy acquired.
We do third-order work at this time. At the time this survey was made, this man's accuracy was 50 lengths per mile, which is 33 feet in a mile. That is what he could be off and still be an accurate survey. The amount he is paid definitely influences the accuracy in which he made the survey.
BY MR. ESSNER:
Q Is it your opinion that the meander line — the traverse which describes the meander line of Banana Creek, that is, the position of it and the known position of the shoreline at this point was the result of an error of inaccurate survey?
A It would be my judgment that it was just an expedient way of getting around this knob that sticks out beyond the rest of the land. (Emphasis added)
Q Would that represent any savings of time and money in making up such an expedient calculation?
A Of course, it would — sure it would.
Q What would be involved in making an exact representation of the shoreline as opposed to drawing a traverse in this fashion?
A Well, now, an exact representation of the shoreline would be almost a hopeless and tedious job as far as the shoreline — you can get closer and closer to the proximation of the exact shoreline. Our survey doesn't represent the exact shoreline.
We have eliminated certain small —
Q In order to get within the tolerances allowed under the Harris survey, what additional time would be involved?
A Two or three days — one field proof, two or three days.
Q To describe the parcel of land that we are referring to here as 3837-A?
A Yes, sir.
R. 100-102. On the basis of this uncontradicted testimony, we conclude that the explanation for Harris' failure to follow the river around the subject tract lies in the type of survey being made and the type of land being surveyed. From all the available evidence, we conclude that he intended to meander the river and to establish the river as the northern boundary on the plat. The negative side of this proposition is that he did not intend to establish the northern boundary of Government Lots 5 and 6 as the northern boundary on the plat, thereby excluding Tract 3837-A.4
Granted that Harris intended to meander the river, the fact remains that those who prepared the plat on which the conveyance was based did not know about the 98-acre tract north of Government Lots 5 and 6. Thus, we pose the final question which is one of law: Was the Government bound by Harris' intention to meander the river or did his failure to indicate the existence of Tract 3837-A in his field notes have the effect of excluding that tract from the conveyance? In his conclusions of law, the trial judge stated what all parties concede to be the general rule that meander lines are not to be treated as boundaries and that when the United States conveys a tract of land by patent referring to an official plat which shows the same bordering on a navigable river, the purchaser takes title up to the water line. Producers Oil Co. v. Hanzen, 1915, 238 U.S. 325, 339, 35 S. Ct. 755, 760, 59 L. Ed. 1330; Horne v. Smith, 1895, 159 U.S. 40, 42, 15 S. Ct. 988, 989, 40 L. Ed. 68. This rule evolved because of the difficulty surveyors had in following the edge or margin of projections of land and all the various sinuosities of the water line. Because of the difficulty of making the description conform exactly to the water line, the sound approach was to consider the lake or stream itself, as a natural object or monument, to be one of the calls of the description or boundary of the premises conveyed. Mitchell v. Smale, 1891, 140 U.S. 406, 413, 11 S. Ct. 819, 821, 35 L. Ed. 442.
In United States v. Lane, 1923, 260 U.S. 662, 43 S. Ct. 236, 67 L. Ed. 448, a case which we believe controls the one at bar, the Court applied the general rule just cited over objection by the Government that certain parcels of land had been omitted by the surveyor, were not shown on the plat, and therefore had not been conveyed by the patent. In 1839, a man named Warren surveyed land near Ferry Lake in Louisiana. The survey reflected that he intended to meander the shore of the lake, but the meander line he described did not in all instances coincide precisely with the water's edge. With respect to the first four cases consolidated for appeal, the Court reached the following conclusion:
* * *., the aggregate of the various parcels lying outside the meander line is about 70 acres, and the aggregate of the various parcels of water included within the meander line is about 44 acres. The facts bring the cases fairly within the rule announced by this court in Mitchell v. Smale, 140 U.S. 406, 11 S. Ct. 819, 840, 35 L. Ed. 442, and not within the exception which was followed in the Jeems Bayou Case [Jeems Bayou Fishing & Hunting Club v. United States, 260 U.S. 561, 43 S. Ct. 205, 67 L. Ed. 402]. So far as the instant cases are concerned, there is nothing in the circumstances to suggest the conclusion that any fraud was committed or palpable mistake made by Warren. At the time of his survey the lands were of such little value, the locality so wild and remote, and the attendant difficulties so great that the expenditure of energy and money necessary to run the lines with minute regard to the sinuosities of the lake would have been quite out of proportion to the gain. We are of opinion that the survey of 1839, except as to the two large tracts just mentioned, is not open to challenge. The precisely accurate survey of 1916-1917 would probably never have been made, but for the greatly increased value of the lands due to the discovery of oil and gas therein.
260 U.S. at 664-665, 43 S. Ct. at 237. With respect to the remaining two cases, the Court also applied the general rule that the water is the boundary though a greater amount of land lay outside of the meander line than in the first four cases:
The Warren survey and plat here indicate an area of about 271 acres in the fractional subdivisions conveyed. The lands added by the new survey constitute a compact body of 97.64 acres — 65.77 acres of which are involved in the Greene Case and 14.13 acres in the Loucks Case — having the outline of a long and rather irregular crescent, with the outer curve next to the water. * * * The evidence justifies the conclusion that in 1839, and especially in time of high water reaching back into the ravines, the establishment of a line precisely coincident with the water's edge would have been a matter of expense and difficulty wholly disproportionate to the then value of the omitted acreage. As in the four cases first discussed, the Warren plat of the survey referred to in the patent represents the lake as the boundary. The survey, taken as a whole, with the exception of the two large tracts already mentioned, follows with a fair degree of accuracy the contour of the lake, and the evident purpose was to include in it all the land to the water's edge. Considering the circumstances in respect to the character and value of the lands, the wildness and remoteness of the region, and the difficulties surrounding the work of the surveyors, the failure to run the lines with more particularity was not unreasonable, and we are constrained to agree with the lower court in holding that the waters of the lake, and not the traverse line, constitute the boundary.
260 U.S. at 666-667, 43 S. Ct. at 237-238.
In the instant case, appellants insist that the plat shows the exact acreage of Government Lots 5 and 6 and that under the cases this amount and no more was conveyed by the patent. While language in some of the cases, especially in Horne v. Smith, supra, may lend itself to this construction, we do not read the cases to say that the designation of acreage controls whenever the plat is divided into sections and lots. In United States v. Lane, for example, the plat showed 271 acres in the fractional subdivisions bordering on the lake, but the Court held that an additional 97.64 acres lying outside the meander line were conveyed by the patent because the lake was intended to be the boundary. Similarly, we hold that 98 acres (Tract 3837-A) in addition to the 98 acres comprising Government Lots 5 and 6 were conveyed by the patent because the Banana River was intended to be the northern boundary. In light of the character and value of the the land and the difficulty of surveying, "the failure to run the lines with more particularity was not unreasonable."5
In Lane, the Court alluded to two large tracts in the area to which the general rule was inapplicable. These tracts were also the subject of litigation. In Jeems Bayou Fishing & Hunting Club v. United States, 1923, 260 U.S. 561, 43 S. Ct. 205, 67 L. Ed. 402, the plat showed a body of water (Ferry Lake) covering an area where there was, in reality, a 500-acre tract of well-timbered land. The Court refused to hold that the water constituted the boundary and that the large tract lying between the water's edge and the boundary line on the plat was therefore a part of the conveyance. As we read the opinion, the general rule of Mitchell v. Smale and United States v. Lane was not applied because a) the size of the tract between the surveyor's boundary line and the water made it improbable that the surveyor intended for the water to be the boundary, b) nothing in his field notes reflected an intention to make the water the boundary, and c) if he intended to make the water the boundary, his failure to cover the large tract in question would be "such gross and palpable error as to constitute in effect a fraud upon the government." 260 U.S. at 564, 43 S. Ct. at 206. In the instant case, the tract not shown on the plat was not so large in relation to the entire amount of land on the plat as to suggest a fraudulent survey. According to Mr. Jobe's undisputed testimony, Harris was justified in not following the river around Billy Joe Point. Moreover, his field notes provide positive evidence that he intended for the river to be the boundary. Taking into account two critical factors — the size of the respective tracts and the intentions of the respective surveyors — we conclude that Jeems Bayou is distinguishable from the case before us. Producers Oil Co. v. Hanzen, supra, the case involving the second of the large tracts referred to in the United States v. Lane, is likewise distinguishable. Again, the Supreme Court held that the tract between the boundary line on the plat and the actual location of the water's edge was not included in the conveyance because the surveyor did not intend for the water to be the boundary at that point. His intention was revealed by a declaration in his field notes that there was a tract of land betwen the traverse line he was describing and the water. In other words, he indicated that he would establish a boundary line short of the water rather than meander the water's edge.
Appellants cite other cases which we find distinguishable for the same reasons as Jeems Bayou and Producers' Oil Co. See Chapman & Dewey Lumber Co. v. St. Francis Levee District, 1914, 232 U.S. 186, 34 S. Ct. 297, 58 L. Ed. 564; Security Land & Exploration Co. v. Burns, 1904, 193 U.S. 167, 24 S. Ct. 425, 48 L. Ed. 662; Niles v. Cedar Point Club, 1899, 175 U.S. 300, 20 S. Ct. 124, 44 L. Ed. 171; Horne v. Smith, supra. Without analyzing each one of them, suffice it to say that we believe United States v. Lane is controlling and that the cases cited by appellants represent an exception to the general rule which is not applicable. The exception is inapplicable because Tract 3837-A is not particularly large in relation to the total acreage on the plat, there is no evidence of such gross error as to constitute a fraudulent survey, and there is positive evidence Harris intended for the river to be the northern boundary. Each case must turn on its facts and must be considered in light of the general rule and the exception to that rule. We do not, as appellants have suggested that we might, hold that any reference on a plat to a body of water will mean as a matter of law that the water is a boundary. Having studied the evidence and the relevant cases, we merely conclude that in this case the river is the boundary and that the judgment of the lower court must accordingly be affirmed.
Since we affirm, it will not be necessary to consider the State's position vis-a-vis the United States. So far as the critical issues are concerned, the two appellants are of one mind
Mr. Jobe testified that Tract 3837-A "is no different in character or nature or vegetation covering than the area lying south of Mr. Harris' government meander line." (R. 97) On cross-examination, he was asked whether any of the tract was fit for cultivation and again said, "It is no different from that south of it." (R. 105)
Jobe testified that he thought Harris intended for the river to be the boundary:
Q Let me put it more directly and more simply. Mr. Jobe, what does the northerly boundary of Lots five and six represent in the government map — what are they supposed to represent according to the instructions?
A I believe I could answer if you would permit me, your first question in that the contract made reference to instructions which are contained in the manual and the manual instructions are to meander as near as practical the mean high waterline in regard to the meander lines.
* * * * *
Q The government map of 1860, Exhibit No. 3. Do you interpret that map to mean that the line drawn along the northerly boundary to be the waterline? A If I had nothing else to go on, I would assume that was the waterline, yes, sir.
Q That is what Harris intended it to be?
A Yes, sir.
R. 91. There is no testimony to the contrary.
We return to the trial judge's finding that Tract 3837-A was "unsurveyed." It was unsurveyed in the sense that Harris did not take the eight courses required to traverse it, i. e., he did not follow the river around it, but the type of survey he was making allowed for some error. He could make a survey roughly corresponding to the course of the river without being precisely accurate. Considering his intention to meander the river and the fact that he had some margin for error, the subject tract was not really "unsurveyed."
We believe the rule is that the designation of acreage on a plat is always a relevant consideration but is never controlling. Chapman & Dewey, Lumber Co. v. St. Francis Levee District, 1914, 232 U.S. 186, 197, 34 S. Ct. 297, 299, 58 L. Ed. 564