delivered the opinion of the court.
The history of the relation of the government of the United States to the claims for lands asserted under rights derived from the Spanish and French governments, prior to the cessions of Louisiana and Florida to our government, as it is found in the treaties, the acts of Congress, and the judicial decisions of the American tribunals, is given very fully and with accuracy in the opinion of this court in the case of The United States v. Lynde (11 Wall. 632), and will be referred to now without repeating it. The necessity and the policy of the act of 1860 are there fully considered. It declares that the registers and receivers of the public land-offices in Florida and Louisiana, within their respective districts, and the recorder of land-titles for the State of Missouri, shall be commissioners to hear the evidence and make report to the Commissioner of the General Land-Office concerning this class of claims. They are directed in their reports to divide the cases into three classes, two of which were to be reported as valid and the third as invalid. The nature and character of these claims, and the evidence on which they are to be held valid or invalid, are fully set out in *417the statute. After the reports of these officers are filed with the Commissioner of the General Land-Office, they are to be subject to the examination of that officer, who is to report thereon directly to Congress. In all cases where he and the local commissioner concur in the rejection of the claim, that action is to be final; but where he concurs with these commissioners in holding a claim valid, he shall report the same to Congress for its action. And in cases where he disapproves the report of the commissioners, he shall in like manner report the whole matter to Congress for final action. It will thus be seen that in all cases brought before any of these officers, under this act, except when the Commissioner of the General Land-Office concurs with them in rejecting the claim, the whole proceeding amounts merely to a report to Congress, and the final action of confirming or rejecting the claim rests with that body.
The eleventh section, however, enacts that in a much more limited class of cases, which it specifically defines, the claimants “ may at their option, instead of submitting their claims to the officer or officers hereinbefore mentioned, proceed by petition in any district court of the United States within whose jurisdiction the lands or any part of the lands claimed may lie, unless such claim comes within the purview of the third section of this act.” It declares that the United States may be made defendant to such a suit, and an appeal allowed prescribes the mode of executing a final decree in favor of the claimant, and provides for other matters. So much of it as excludes claims coming within the purview of the third section evidently has reference to the proviso of that section, that no case shall be reported favorably by the commissioners which has already been twice rejected on its merits by previous boards, or has been rejected as fraudulent, or as having been procured or maintained by fraudulent or improper means.
The difference in these two modes of procedure, and in the results which followed them, are obvious and important. The first, as already observed, is merely a mode of placing before Congress the result of an investigation by the local commissioner, and the Commissioner of the General Land-Office, with their opinion on the merits of the claim. On these reports Congress either rejects or confirms the claim, as it may think *418right. Until such action by Congress, nothing is concluded; and if it fails to act, the previous inquiry amounts to nothing.
The suit in the District Court, on the other hand, has all the elements of any other judicial proceeding, among which are the conclusiveness of the judgment on both parties, and the right to an appeal to this court for final decision. Considering the more valuable results which may be obtained in the courts, and the better-defined course of procedure there, it is not strange that parties who have faith in the validity of their claims should prefer that tribunal.
But Congress did not intend to refer all the cases embraced in the act to the courts, at the option of the claimant. It was only claims of a class defined by the eleventh section of the act, which the claimant might bring either before the court or before the commissioner, at his election. If the case before us does not belong to this class, the court did right in dismissing the petition, whatever may be its merits, and though it may be a case which, if brought before a commissioner, would be entitled to a favorable report.
We must, therefore, examine the case in the light of the provisions of the eleventh section, which defines this class in these words: —
“ Any case of such a claim to lands as is hereinbefore in the first section of this act mentioned, where the lands claimed have not been in possession of and cultivated by the original claimant or claimants, or those holding title under him or them, for the period of twenty years aforesaid, and where such lands are claimed by complete grant or concession, or order of survey duly executed, or by other mode of investiture of the title thereto in the original claimant or claimants, by separation thereof from the mass of the public domain, either by actual survey or definition of fixed natural and ascertainable boundaries or initial points, courses, and distances, by the competent authority prior to the cession to the United States of the territory in which said lands were included, or where such title was created and perfected during the period while the foreign governments from which it emanated claimed sovereignty over, or had the actual possession of, such territory.”
A careful examination shows three distinguishing elements necessary to a suit in the court: —
*4191. The claimant or those under whom he holds must have been out of possession for twenty years or more.
2. The land must be claimed by a complete grant or concession, or order of survey duly executed, or other mode of investiture of the title in the original claimant by separation from the mass of the public domains, either by actual survey or defined fixed natural boundaries or initial points and courses and distances, by the competent authority, prior to the cession to the United States.
3. Where such title was created and perfected during the period of the actual possession of the prior government under which the claim is asserted.
This is substantially an action of ejectment, with the bar of the Statute of Limitations removed, the United States having a constructive possession for the defendant.
The title must be complete under the foreign government. The land must have been identified by an actual survey with metes and bounds, or the description in the grant must be such that judgment can be rendered with precision by such metes and bounds, natural or otherwise.
There must be nothing left to doubt or discretion in its location. If there is no previous actual survey which a surveyor can follow and find each line and its length, there must be such a description of natural objects for boundaries that he can do the same thing de novo. The separation from the public domain must not be a new or conjectural separation, with any element of discretion or uncertainty.
The right to sue here given is not on an inchoate or imperfect title. It is not on a perfected grant for an unknown locar tion, or for a given quantity within defined out-boundaries. All these are left to be pursued, if at all, before the commissioners appointed by the statute. They could pass upon the equities arising from imperfect or incomplete grants. An order of survey was sufficient before them, if otherwise sustained by proof. Permission to settle on the land, or any other written evidence of title emanating from the foreign government prior to the cession. This required no completed title, no actual survey, no twenty years out of possession, no prior segregation from the public domain. In all this class of cases, Congress, which *420reserved the right to decide, only required evidence of some equitable claim arising under the former government, on which, it could make an intelligent decision.
But in the cases brought before the courts, while removing the bar of the lapse of time, and the want of a defendant in possession, and the defence of a better title by patent from the United States, the act still requires a title completed under the foreign government, evidenced by written grant, actual survey, or investiture of possession, and, in short, evidence of a title on which recovery of possession could be had when these defences were out of the way. This view is confirmed by the provision that the petitioner must have been out of possession for twenty years. The only reason that occurs to us for this is, that having the superior legal title, he could recover from any one in adverse possession without the aid of the statute, where he was not bound by twenty years’ limitation.
Does the case before us come within this class ?
There was no actual survey. The order of survey made by Governor Carondelet is very indefinite. It is thus translated in the record : —
“ 11th June, 1793, to Captain Don Joseph Valliere, in the District of Arkansas, a tract of land situated on the White River, extending from the rivers Norte Grande and Cibolos to the source of said White River, ten leagues in depth.
“ Baron de Carondelet.”
On the strength of this order, Trudeau, the surveyor-general, proceeded to make what he calls in his certificate of survey “ a figurative plan ” by conjecture, and from this gives a certificate of survey. It appears by the paper called a grant and signed by Carondelet that this plan was made by his order because it was impossible for the royal surveyor to make an actual survey at the time.
Based upon this figurative plan, the concession or grant makes an attempt to describe the land granted by certain natural objects, and some general but not specific directions as to the courses and distances. It does not appear that any actual survey has ever been made locating this grant. It does not appear that any attempt has ever been made to do it. We *421have in the record copies of Trudeau’s sketch. We have a copy of the official map of the surveys of the land into congressional subdivisions, made for the purpose of selling these lands, which have been extended over all the area in which this grant could possibly be found; and we have a map of the State of Arkansas, with county and township subdivisions ; and in both these latter the general coui’se of the White River, its branches and affluents, are laid down.
On this latter map we have what Mr. Day, a civil engineer, swears to be a correct location of this grant according to boundaries given in Carondelet’s cession. This was not made by any actual survey, but simply taking the sectional map of the State of Arkansas, Mr. Day has made lines on it, which he declares to be a location, on that map, of Yalliere’s grant. He does this by assuming a point in township 13 north, range 25 west, in Arkansas, to be the origin of the White River, and proceeding directly south from this point ten leagues, or 37£ miles, he makes a series of arbitrary lines, with a corresponding number of angles and changes of course, tending first northwest, and then northeast, and then southeast, until he reaches the Great North Fork of said river. He then descends said fork until it intersects the river, descends the main river until he reaches Buffalo Fork, ascends Buffalo Fork until he comes near the initial point or source of White River, and then makes a straight and arbitrary line southwest to the beginning. As regards this survey, the straight lines and the changing courses and distances are wholly arbitrary and artificial, having no natural objects to establish them, and nothing in the descriptive language of the grant. They are intended to be the conjectural or average distances of ten leagues from the White River. That is to say, in a distance of nearly three hundred miles on one side of White River, in order to ascertain definitely what lands are within ten leagues of that river, — one of the most tortuous ever known, — the surveyor makes six new departures and courses, and, running these by straight lines, declares that he has solved the problem and made an accurate survey.
But let us compare this survey with the calls of the grant. The latter describes the land as “ situated on both banks of White River, ten leagues on both banks, beginning at the origin *422of the most western branch or source of the White River, and running southwest ten leagues, descending thence on the South by parallel line with White River, at the distance of ten leagues, until it intersects Buffalo River at a point ten leagues in a direct line with White River, from thence descending the Buffalo River to its confluence with White River; following this as far as the mouth of the Great North Fork of the White River, up the same to a point ten leagues in a direct line from its mouth, from thence ascending the White River to the north in a westerly direction, ten leagues from the same, as far as its source, which will better be seen on the figurative plan,” &c.
Assuming that Day’s survey has located the original source of White River as the initial point correctly, the first call in the grant is southwest ten leagues. Mr. Day’s line is ten leagues directly south; the next departure in the grant is descending thus on the south by parallel with the White River at the distance of ten leagues, until it intersects the Buffalo River at a point ten leagues in a direct line with White River. Here Mr. Day utterly disregards the call, makes a due west line, taking him directly away from the Buffalo River, and making his artificial courses and distances nearly three hundred miles, not on the south, but on the west and north, of White River, and never gets to Buffalo River until he has run the reverse course of the call, and meets it near the last of his survey at its junction with White River. The reason of this is obvious. The call in the grant is an impossible call. The Buffalo River is not in the direction supposed by Trudeau and Carondelet, and the source of White River is not where it is supposed to be.
The next call in the grant is to descend the Buffalo to its confluence with White River. But the Buffalo would never be reached by the call of the grant. In short, looking at the calls for material objects, courses of streams, and distances, that which might have been predicted occurred. In attempting to make a grant described by rivers of whose courses and location they were ignorant, by given distances which could not be made to conform to the natural objects, making a grant of over seven millions of acres of land by specific boundaries of which they knew nothing, they made a total failure, and gave no *423description by which any surveyor could, without the aid of a lively imagination, make any location.
This is clearly manifest by a comparison of Trudeau’s plan with Day’s location, and with the actual locality and course of the streams as they are now ascertained.
Trudeau’s plan and the calls of the grant make the initial point and source of White River in the northeast corner of the plat; Day makes the source of the river and the initial point in the middle south part of his survey. Trudeau runs a waving line in a southeastern direction to Buffalo River, where he supposed it to be; Day runs in a reverse direction northwest, until he meets the North Fork, and comes down it.
Trudeau was mistaken if the source of the river is where Day locates it. But this destroys all Trudeau’s plan, and locates the grant in a very different place from where he and Carondelet intended it to be, and where it can never be reached by any survey following the description of the grant.
But on what evidence Mr. Day relies to fix the source of the river, the beginning point of his location, is unknown. He did not go on the ground or trace the stream. He merely takes the map of Arkansas, and say's, here on this map I find the origin of the river to be a point in township 13 north, range 25 west, in Madison County.
Whether this map gave the origin of the most western branch of that river correctly is wholly uncertain. How far a surveyor must pursue such a branch or stream to find the fountain from which it flows is left in the dark. If Mr. Day had gone on the ground, ordered to make the survey under oath, he might have felt bound to locate this point many miles from where he finds it on the map. It is almost absurd to suppose .that in an ordinary traveller’s map of a State, made to be folded into a pocket-case, any reliance can be placed on its location of the source of a stream which would justify its acceptance as a warrant for locating with precision a grant of over seven millions of acres of land. The combined exhibits E and F, which are certified copies of the official surveys of the United States, call this most western branch Buffalo Fork, and do not locate the origin of this western branch within thirty miles of the point which the Arkansas map does, and where Mr. Day does.
*424We are of opinion that for want of any actual survey at the time the grant was made, or at any other time, by the Spanish government, for want of any other separation of the land granted from the mass of the public domain, and for want of any description of the land granted in the instrument of cession, or order of survey, by which the land can be surveyed and identiffed, the claim does not come within the eleventh section of the act of 1860, and that the District Court properly rejected it.
Decree affirmed.