delivered the opinion of the Court.

These suits involve claims of title on the part of the United States, hereinafter called the plaintiff, to various parcels of land lying along the border of Ferry Lake, a navigable body of water in Caddo Parish, Louisiana. Answering these claims, the defendants in the respective cases averred that plaintiff, long before the bringing of the suits, had conveyed by patents to private persons certain fractional subdivisions bordering on the lake; that these fractions were represented on the official plat of the government survey, made by one Warren in 1839 and duly approved and filed, as bounded on the lake side by the waters of the lake; that in each of the cases the land in controversy was a small tract, lying along the edge of the lake and constituting.part of the particular fractional subdivision so conveyed; and that consequently, plaintiff had divested itself of whatever title it originally had.

Certain alternative defenses, based upon the alleged ownership of the lands by the State of Louisiana, were pleaded, but, in view of the conclusions we have reached, it is riot necessary to consider them. The District Court entered decrees for the plaintiff which the Circuit. Court of Appeals reversed (274 Fed. 145, 290), and the cases are here upon appeal except the Loucks case,, which comes-on certiorari. The foregoing averments of fact contained in the answers are established by the record.

*664 In 1916-1917, nearly eighty years after the Warren •survey — the lands in the meantime having become valuable for their deposits of oil and gas — a new survey was made under the direction of the General Land Office. This survey shows that the line run'by* Warren, purport-' ing to meander the shore of the lake, did not in all instances'coincide precisely with the water’s edge. In the four cases first named in the title, the parcels of land lying between the meander line and the lake are of small extent. The meander line throughout its length approximately conforms to the' sinuosities of the shore, sometimes,' however, running for short distances inland and sometimes for short distances into and through the water.In the first mentioned suit the Warren survey indicates a fractional subdivision containing 26.80 acres, the new survey adds 5.67 acres. In the second suit the Warren survey indicates 23 acres, the new súrvey adds 12.72 acres. In the third suit the Warren survey indicates 155 acres, the new survey adds 27.87 acres.. In the fourth suit the Warren survey indicates 114.80 acres, and the new survey adds 11.49 acres.

The lands in question are all in township twenty; the first mentioned being in section 3, the second in section 10, the third in section 13 and the last in section 24. Following the meander line of the Warren survey the distance from the first of these tracts to the last is about five miles. Leaving out of consideration the large tract involved in Producers Oil Co. v. Hanzen, 238 U. S. 325, and the large tract involved in Jeems Bayou Fishing & Hunting Club v. United States, ante, 561, the aggregate of the various parcels lying outside the meander line is about 70 acres, and the aggregate of the various areas 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, and not within the exception which was followed in the deems *665 Bayou Case. 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. It is unnecessary to do more than quote the language of this Court in Mitchell v. Smale, supra, 413:

“The pretence for making such surveys, arising from the fact .that strips and tongues of land are found to project into the water beyond the meander line run for the purpose of getting its general contour, and of measuring the quantity to be paid for, will always exist, .since .such irregular projections do always, or in most cases, exist. The difficulty of following the edge or margin of such projections, and all the various sinuosities of the water line, is the very occasion; and cause of running the meander line, which by its exclusions and inclusions of such irregularities of contour produces an ¿verage result closely approximating to the truth as to the' quantity of upland contained in the fractional lots bordering on the lake or stream-. The official plat made from such survey does ■not show the meander line, but shows the general form of the lake deduced therefrom, and thé surrounding fractional lots adjoining and bordering on the same. The patents when issued refer to this plat for .identification of the lots conveyed, and áre equivalent to and have the legal effect of a declaration'' that they exténd to and are *666 bounded by- the lake or • stream. Such lake or ' stream itself, as a natural object or monument, is virtually and truly one of the calls of the description or boundary of the premises conveyed; and all the legal consequences of such a boundary, in the matter of riparian rights and title to land under water, regularly follow.”

While the facts are somewhat different and the extent of the omission.somewhat greater, we think the general rule should, likewise; be applied in the remaining two cases. The Warren survey and plat here indicate an area of about 271 acres in theiractional subdivisions conveyed. The lands added by the new survey constitute a compact bo'cly of 97.64 acres—65.77 acres of which are involved in the Greene case and "14.13 acres in the Loueks case—having the outline of a long and rather irregular crescent, with the outer curve next to the water. The inner boundary, running between the two points of the eres--cent, is made up of a series of straight lines, with intervening ¿ngles conforming to the outlying curve only to-a-roughly approximate degree. The length of the tract is nearly 4,000 feet and the extreme width -about 1,200 feet. Running back from the shore there are numerous ravines, creating, in and along the outer rim of the tract, a series of alternating points and indentations. The evidence justifies the conclusion that in 1839, and especially in time óf 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 wdiolly disproportionate to the then value of the omitted acreage. Asjn 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 two large tracts already mentioned, follows wi;h a fair degree of accuracy the contour of the lake and the evident purpose vras to include in it ¿11 the land to the water’s edge. Considering *667 ■the circumstances in respect of 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.

The decrees of the Circuit Court of Appeals in all the cases are, therefore,

Affirmed.