The deputy surveyor, Rice, plainly *706limited the fractional quarter sections to the territory separated by the meander-line from the “ impassable water and marsh,” and surveyed “ Sandy ” and “ Crane ” as separate and distinct “islands.” The surveyor-general, Lytle, so treated fractional sections and islands, in making the statutory plat. The patents to Margaret Bailey conveyed to her each fractional section “ according to the official plat of the survey of said lands returned to the General Land Office by the surveyor-general.” James, at the trial below, put in evidence a plat, which the commissioner of the General Land Office, on April 16, 1851, certified to be a “ true and literal exemplification of the plats on file in this office.” This copy showed “ Cedar” and “ Crane ” as islands. On it no township or section line crossed either, except that the north and east lines of section 4, township 10 south, range 10 east, are shown. -This north line crossed a small island west of “ Crane,” while th‘e east line crossed the west end of “ Crane.” The lines of no other fractional section, lying south of Sandy and Crane islands, extend into the marsh at all. About one fourth of Sandy Island (the western part), is shown as an island. The remainder appears, some of it as a straight black line extending southeasterly from said western portion, and part as the vevy small island above referred to as crossed by the north line of said section 4. Because of this certified plat, James claims that, under power given by the act of July 4, 1836 (Land Laws, vol. 1, p. 552), the commissioner had disapproved so much of what had been done by the surveyor-general and his deputy as treated Sandy and Crane as islands, and therefore, they became and remained parts of the fractional sections.

Howell put in evidence certified copies of the surveyor-general’s plats of the three townships, and a certified copy of a diagram of the islands, made pursuant to the act of January 22, 1853. This was certified July 30, 1875. All these showed the islands as separately surveyed and computed. They were not included within any surveyed section lines, although on each township plat appeared dotted *707projections of the section lines across water, marsh and. islands. It is evident that the copy certified April 16th, 1851, is unreliable. It shows townships 9 and 10 south of range 10 east, as lying west of townships 9 and 10 south of range 9 east. If the commissioner had disapproved of the separate survey and platting of the islands, his approved plat would not show all of Cedar and Crane, and part of Sandjq as islands not included in sections. Instead of entirely omitting the dotted lines of Lytle, crossing water, islands and marsh, he would have converted them into regular section lines. An inspection of the certified copy, we think, establishes that this claim of James is unfounded.

The lines and corners, as actually surveyed and marked, bound the fractional sections by what is called the “ impassable water and marsh,” and not by the main body of the lake. The “ meander ” line along the southerly border of the marsh Was, in fact, intended to be the boundary line of the fractional sections. It is clear that the United States Land Office did not treat the islands as parcel of those fractional sections; it did treat them as islands surveyed, platted and to be sold as such. Margaret Bailey’s patents by express terms, grant “according” to the surveyor-general’s plats. The patents must be so construed as to give effect to their true intent and meaning, but the rule favoring grantee as against grantor does not apply to them, because the vendor was the public, the nation.

It is urged for James, that, under the act of 1796, it was unlawful to bound the fractional sections by the .marsh; that by force of that act, each of them extends across marsh and island to the main lake; that the statute authorized no boundaries for fractional sections except (1) straight north and south, or east and west lines; (2) “the line of the late Indian purchase; ” (3) “ tracts of land heretofore surveyed and patented;” or (4) “the course of navigable rivers';” and, therefore, wherever the lines of one of these fractional sections, if projected so as to include a square mile, do not meet either of these legal boundaries until they enter the *708main lake, all of that square mile between said lines and the lake are parcel o'f thé section, and pass to its patentee.

The act of 1796 is the instruction given by the government, as principal, to its agents in the land office. If construed literally and strictly, lake Erie itself could not be a legal boundary. Although in the mind of- a geographer it may be considered as part of the river St. Lawrence, it'is in legal terms, and in popular parlance, not a river, but a lake. But there is no dciubt that it is. a legal boundary under the true intent and meaning of the act of 1796. 'The waters of this lake are added to by the streams that drain the lands around it. The contributions of , such streams, mingled with other water of the lake,' fill the straits between Cedar, Sandy and Crane islands, and the space platted as “impassable water and marsh.” Even at low water in the lake, there is six inches of water in this marsh. Witnesses speak of a chain of ponds. These are merely spots in the marsh where, because of the depth, or for some other reason, vegetation has not been able to find support. The water is under the vegetation that covers the so called marsh, as well as in the so called ponds. The islands are in fact surrounded by water, although' vegetation of a certain sort grows thickly in much of that water. The meandered line along the southerly edge of the so called marsh, was in fact along a shore of the lake. In order to form a statutory boundary, the navigable river need not be actually navigable at every point. Lake Erie as a whole is navigable. For that reason every part of its shore, under the act, might be 'trsated by the surveyor as a legal boundary for fractional townships and sections. In so treating so much of the lake as lay behind these islands, Rice and Lytle did not exceed their authority. According to’the evidence put in by James, as well’ as by Howell, it appears that even at a low stage <jf the lake there is six inches depth of water in this inarsh, and a greater depth in the straits. Moreover, the act of 1805 expressly provides that “ the boundary lines actually run 'and marked in the surveys returned by the surveyor-general, shall be established as the proper boundary lines of the sections *709or subdivisions for tohich they were intended, and the length of such lines as returned shall be held and considered as the true length thereof.” Including the “ meander,” all the lines o£ Margaret Bailey’s fractional sections were actually run; marked and returned. Hence the provision touching lines not so run and marked, does not apply to her lands.

On behalf of James, her assignee, R. R. Co. v. Schurmeir, 7 Wall., 272, is cited in support of his claim, that the “meander” line, along the southern edge of the marsh, was run, not as a boundary for the fractional sections, but “ for the purpose of defining the sinuosities of the banks of the marsh, and as the means of ascertaining the quantity of land in the fraction which was to be paid for by the purchaser.” In that ease the court so held, but in stating the facts it said: “ In the government survey, no mention of, or reference to, this bar or island, was in any way made in the field notes, plat or map. The fractional parcel, as already said, was represented as lying immediately upon, and bounded by, the Mississippi river.” The opposing claimant based his title on a second survey, made after the patent had been issued under the original survey. Clifford, J., in the opinion, after reciting the facts and quoting the statutes, said: “Express decision of the Supreme Court of the state was, that the river, in this case, and not the meander line, is the west boundary of the lot, and in that conclusion of the state court we entirely concur.” This decision was made at December term, A. d. 1868, Miller, J., concurring. Sitting in the circuit court for Wisconsin in April, 1865, the latter judge tried Granger v. Swart, Woolworth C. C. Rep., 89. He said to the jury: “ The patents .....under which the defendant claims, do not pass the title to the premises in question, unless, at the date of the entries on which they issued, the Rock river, where it is called a river, and lake Koshkonong, where it is called a lake, extended to and bordered upon the meandered line which constitutes the boundary of the lands described in the patents. In other words, if, between the meandered line, which by the government survey was made one of the *710boundaries of the land sold to Walker, and the 'bank of Rock river and shore of lake Koshkonong, there was at that time a body of swamp, or waste land, or flats, on which timber and grass grew, and horses and cattle could feed, and hay be cut, then the patents to Walker did not cover this land, but was confined to the actual limits of said meandered line.”

If the claim of counsel lor James be true, this charge was bad law; the surveyor could not, by the meander line, deprive the fractional section of its right to go as far as to the river or lake.

These two cases do not conflict. In each case it seems to us the same principle was applied. In the absence of plain indication of an intent to make the “ meander line ” the actual boundary of the fractional section," it must be treated as run for the purpose of ascertaining the quantity to be paid for by the purchaser, and the section extends, with the lines of its square mile, to the actual stream or lake. In the case before us, the intent to limit the fractional sections by the marsh, and to exclude the islands, is unmistakable. Hence, even if the evidence had shown that the marsh contained far more earth than water, and that hay could be cut, or cattle pastured, thereon, James could not hold the islands beyond the marsh.

But, as already stated, that evidence established the permanent presence of water between the islands and James’ meander line; that this water was maintained there by the lake and, for the purpose of constituting a boundary under the act of 1796, was a part of the lake. It is true the channel between the meander line and the islands was not navigable, but that, in our view, is immaterial. The statute does not read “ navigable water,” or “ navigable channel,” but “ navigable rivers.” Hence, where a river itself is navigable, the government surveyors may treat any part of such river as a boundary, as well a part that will not permit the passage of a boat, as those deep enough to float the largest vessel. .

Rice and Lytle treated the bank along the southern edge *711of the so called “impassable marsh and water” as the shore of lake Erie; the marsh as a part of the lake, and Cedar, Sandy and Crane, as islands in that lake. It does not appear that the commissioner of the land office disapproved, their action. We think their action was legal, and the judgments complained of are

Affirmed.