ALDRICH, District Judge.
The Libbys and Dudley F. Leavitt were owners in common of the township of Odell, in the county of Coos, and on the 4th day of October, 1882, in the words following, conveyed to Sumner W. Thompson “ten thousand acres of land in
Odell, in said county of Coos, to be surveyed off and bounded north by Irving’s location, south by Stark, and east Dummer and Millsfield, and to extend far enough west to a line to be surveyed parallel with the line on Dummer and Millsfield to include said ten thousand acres.” The term, “off the easterly side of said township,” used in the conveyance from the Libbys to the Weston Lumber Company, and which counsel for the Weston Lumber Company strenuously urge, was not employed in the earlier deed from the Libbys and Leav*244itt to Thompson, under which the rights of the parties must be determined, and can, therefore, have no bearing upon the question of construction presented by this controversy. The Glen Manufacturing Company is, and was at the time of the alleged trespasses, owner of the interest conveyed to Thompson, and the Weston Lumber Company was the owner of the remaining portion of Odell. At the time of the conveyance to Thompson, the township of Odell and the outlying territory was a remote wilderness, and its chief value consisted in standing timber, and the parties to the deed had comparatively little, if any, actual knowledge thereof, but acted in respect to the conveyance mainly upon information gathered from surveyors, and from maps and plans of Odell and the surrounding townships. Putting ourselves as near as may be in the situation of the parties to the deed, and considering all the evidence relating to the transaction, we find, as a matter of fact, that the parties understood at the time that the 10.000 acres would not extend westerly of a line drawn parallel with the Dummer and Millsfield line, from Stark to the southwest comer of Irving’s location. It therefore follows, conclusively, of course, that the parties did not understand, and did not intend, that any part of the 10,000 acres were to extend further north than the south line of Irving’s location. The term “ten thousand acres,” however, being absolute and controlling, and used in connection with the later expression in the description, “to extend far enough west to a line to be surveyed parallel with the line of Dummer and Millsfield to include said ten thousand acres,” if it turns out, as the parties now concede, that there is not 10,000 acres between Irving’s location on the north, Dummer and Millsfield on the east, Stark on the south, and a line from thence drawn to the southwest comer of Irving’s location, then the deed must be accepted as calling for a line parallel with Dummer and Millsfield, drawn from the north line of Stark from a point far enough west to include 10,000 acres when intersecting a line extended from the southwest corner of Irving’s location on the course of the south line thereof. This follows from the fact that Irving’s location is expressly named as the north bound of the land conveyed; and, finding the intention of the parties from the deed and the surroundings, we must say that the parties did not intend to go further north than Irving’s location, and did intend to go far enough west to include 10.000 acres; and, it being necessary, as is conceded, in order to include 10,000 acres, to go further west than a line drawn from Stark to the southwest corner of Irving’s location, then, to give effect to the intention of the parties, we find that a line should be supplied from the southwest corner of Irving’s location, drawn on the course of the south line thereof, until it intersects the necessary west line of the 10.000 acres extended from the north line of Stark. Ascertaining the intention of the parties as a question of law and fact under the New Hampshire rule, we find this to be the effect of the deed, and that the necessary line may well be supplied under the doctrine of Land Co. v. Saunders, 103 U. S. 316, 322; Winnipisiogee Paper Co. v. New Hampshire Land Co., 59 Fed. 542, 547.
The deed now under consideration conveys a parcel of land to be surveyed off from a larger tract, and by its express terms provides for *245a survey to be subsequently made, and tbe important question in controversy is whether the contemplated divisional line has been established. Being aided in the construction of this provision by the acts of the parties under the deed, we find that the parties understood that the grantor should make the survey, and establish a dividing line which was to become a boundary. The Glen Manufacturing Company, the defendant in one case and plaintiff in the other, relies upon a line which is called the “Williams Line,” and insists that it was established as an “agreed line” under the deed, and, if it is not an “agreed line” in the sense in which that term is understood, that it was a survey authorized, made, and acquiesced in by the grantors in the Thompson deed, and therefore controlling as a boundary line; while, on the other hand, the Weston Lumber Company, plaintiff in one case and defendant in the other, relies upon a line called the “Gile Line,” which leaves the Stark town line at a point considerably more than 100 rods east of the starting point of the Williams line, and extending parallel with the Dummer and Millsfield line northerly to Columbia line, passing to the westward of the southwest corner of Irving’s location, to a point on Columbia line about three-fourths of a mile north of the south line of Irving’s location. The Weston Lumber Company insists upon this line as one inclosing 10,000 acres from actual measurements, and as á line based upon the survey contemplated by the deed, and, in answering the calls thereof, one which becomes an established boundary line between the parties. Speaking generally, the territory in dispute is a tract of land considerably more than 120 rods in width and something more than 9 miles in length, and it is sufficient, for the purposes of the questions now presented, to say that both parties have cut and removed large quantities of timber from the disputed territory.
We will first give our attention to the line contended for by the Glen Manufacturing Company, and in respect to this we find that in October after the conveyance Thompson called for a survey, and, after conference between Leavitt and George W. Libby (who was authorized to act for the other Libbys), agreed that Williams, who was to be paid, and was subsequently paid, by the Libbys, should establish the divisional line contemplated by the deed, and that the 10,000 acres should be ascertained by a calculation based upon the lines of the outlying territory already marked, and the courses and distances as shown by plans thereof and upon one Bucknam’s survey, and measurements of the west line of Dummer and Millsfield, which was the east line of Odell. It was understood that Williams, after making his calculations as to how far it would be necessary to go west on the Stark line to include the 10,000 acres, should run north to the south line of Irving’s location. Soon after the conference between Leavitt and Libby, this means of ascertaining the bound was submitted to Thompson, who approved of it, and furnished men to assist in measuring and marking the line. Williams, in making his calculations on paper, taking the town lines and measurements as a basis, found that Irving’s location did not extend far enough west to form a complete northern boundary of the 10,000 acres, and with Leavitt’s assent (though not known to the Libbys at the time), and for the pur*246pose of plotting the 10,000 acres on a plan, drew an imaginary line from the southwest corner of Irving’s location parallel with the Stark line to intersect his supposed west line, which was to run north from the Stark line. This tentative scheme for ascertaining the 10,000 acres, and for establishing a boundary thereof, was submitted to Thompson by Williams as one assented to by Leavitt; and, Thomp-, son approving, Williams proceeded to lay off the 10,000 acres in accordance therewith by running a line north from Stark, and marking the trees in the line, extending this line parallel with the line of Dummer and Millsfield, and, upon measurements, to a point from which he supposed in running easterly parallel with the north line of Stark he would connect with the southwest corner of Irving’s location, thereby closing in the 10,000 acres. At this point the work was interrupted by a storm and lack of provisions, and Williams, desiring to ascertain definitely by further surveys and measurements whether he was, as a matter of fact, rather than supposition, at a point from which he could close in upon the southwest corner of Irving’s location, left the work unfinished. He submitted what he had done to Thompson, who said it was all right, so far as he was concerned, as it would be a long time before he would get up there for timber, and that they could close it up at some other time. What had been done was also submitted to Leavitt and to George W. Libby, who said it was all right, as it was so far up in the wilderness beyond any line where they were reaching with logging operations that there was no hurry about it. As a matter of fact, as subsequent investigations disclosed, when Williams left his work in 1882, he was not opposite the southwest corner of Irving’s location, but considerably north of the point from which he could connect with such comer in accordance with the scheme under which he was operating. . We are requested to find, and do find, that Williams, shortly after he came out of the woods, put a line upon a plan running on paper from the southwest comer of Irving’s location, parallel with Stark line, to his marked line extending northerly from the town of Stark, and that he sent a copy of this plan to Thompson, and gave a copy to Leavitt; but we do not find that this plan, or a copy thereof, came into the possession of the Libbys. In 1890, Williams, under the direction of parties holding the Thompson interest, ran a line from his unfinished line of 1882 parallel with Stark line to the southwest comer of Irving’s location, abandoning the northerly portion of the line run by him in 1882, or so much thereof as extended north of Irving’s location. It is not claimed, however, that this work influences the standing of his line of 1882, for the reason that it was after the controversy had arisen, and that he proceeded ex parte, and unauthorized, in what he did in 1890, so far as concerns the supposed agreed line.
We have given the theory of the Glen Manufacturing Company as to an agreed line our most careful consideration, having in view, of course, the well-settled New Hampshire doctrine that in cases of disputed and uncertain boundaries an executed parol agreement establishing a dividing line between adjoining owners of real estate is conclusive upon the parties and all persons claiming under them. Applying this rule as to the binding force of executed parol agreements *247to a division of land, even in cases where the necessary New Hampshire conditions of dispute or uncertainty exist, speaking generally, is somewhat exceptional, and we think a party relying upon a supposed agreed line which shall limit or extend the operation of a deed or grant of real estate should at least be required to establish a completely executed agreement and line by clear and unequivocal evidence, not only that the agreement was made, but that it was fully executed. The evidence in this case, we think, does not warrant such a finding. The work was left unfinished, and with considerable uncertainty as to the relation which the point of ending sustained to the southwest corner of Irving’s location, the ultimate bound with which the surveyor was to connect his work. The fact that the parties left their attempt to make an agreed line unfinished cannot be supplied by considerations of equity. It remains unfinished and incomplete, and, therefore, like any other incomplete agreement, is not binding. The same views and the same findings dispose of the claim of the Glen Manufacturing Company in respect to this work as a completed survey under the terms of the deed, and we find as a matter of fact that the Williams line is not a controlling divisional line between the parties.
Now, as to the Gile line. This line was ex parte, but should not fail for that reason, provided it answers the calls of the deed. In the first place, we find that the Gile line was not laid upon the ground in accordance with the intention of the parties, as we gather such intention from the deed, aided by the situation and conduct of the parties. We have already found, in connection with the deed in question, that the parties did not contemplate going further north than the south line of Irving’s location. The Gile line, therefore, is such a departure from the manifest intention of the parties that it must be found not to be in accordance with the calls of the deed.
Now, as to the merit of the measurements and calculation involved in Gile’s survey. The deed contemplates a survey, and the character of the survey is necessarily involved in the construction of the deed which must be made with reference to the situation as disclosed by the evidence, including the character of the territory and the knowledge of the parties. Having in view these considerations, and determining this question as a question of law and fact, we think the survey intended calls for allowances for losses, which experience and observation show, inevitably and necessarily result in measuring wild lands under the conditions disclosed by the evidence in this case. Such allowances are, from the nature of the undertaking, somewhat discretionary, but are nevertheless necessary, in order to get the number of acres called for by the deed. At least a large part of the Gile measurements were made, not on a compass line, but on the lines of marked trees, which were not infrequently a little distance from the true line, thus involving short tangents and angles, and with an effort to level the chain over the uneven and precipitous surfaces without making the usual and necessary allowances; and, as a consequence, we think some loss of land must have resulted. One element of loss would come, as it would seem, from taking up too much of the required distance by measuring on the succession of short tangents *248■which resulted from angles in his line of measurement produced by running from tree to tree on varying courses, thus getting over too little land. In other words, a straight, horizontal line, which is the true and controlling line, would reach further, and include more land, than a line run under the method adopted by Gile; or, still again, to use a simple illustration, the Gile method was somewhat like the measurement of a chain where you get its length by measuring over and around angles which take up a certain per cent, of its reaching capacity, whereas, if the chain were to be used in measuring land or other surfaces, it should be straightened, and applied extento. Another element of loss would result from the failure to make necessary allowances over the uneven surfaces, for if you were to reduce the uneven surfaces to a level by cuts and fills, and measure upon a straight line on the level surface, the required distance would carry you further than the Gile measurement. Any deviation of the tape or chain from the true line, whether above or below the horizontal, or lateral from the straight line, however slight, in a degree retards the movement over the ground in the direction of the point on the face of the earth which the horizontal line calls for in order to give the area required by a calculation based upon the number of square rods necessary to make an acre of land. In measuring land under open, free, level, and unobstructed conditions, where the sight is unerring, the method of leveling the chain would probably be the best method for ascertaining the length of the horizontal line required. This is doubtless a true method under any conditions where deviations from a true horizontal line can be avoided; but our findings do not relate to such a situation or such conditions. The problem is quite different when the length of a horizontal line is to be ascertained over mountains and mountain gorges, where the chainman is constantly going over rising and descending ground, where the unevenness is sometimes slight and almost imperceptible, and again abrupt, where almost perpendicular ascents are to be overcome, to be succeeded by precipitous declivities in a thickly-wooded and rocky country, where the view is obstructed, and at least imperfect. Under such conditions, it is almost inevitable that the chain is rarely, if ever, at a true horizontal, and error is therefore insidiously, but constantly, accumulating, and, as a rule, is one way, and that way is in the direction of too much length in the line of measurement and too little land. At least the evidence and experience demonstrate that the inconvenience, difficulty, and oftentimes impossibility, of bringing the chain to a proper level, promote error, and that the resulting balance of error is unquestionably and decidedly in favor of the loss of land. If a line 10 miles in length were laid upon a level land surface by careful measurements, and two end lines were laid at right angles therewith on level surfaces, and connected by a second side line over an uneven mountain surface, the most careful and painstaking measurement with attempts to level the chain with the eye over the mountains involved in the second line would ordinarily, and perhaps universally, develop a longer line than the side line on the level ground. The horizontal side lines would, of course, be exactly the same, but the faulty method of measuring over the mountains would develop a fictitious length. The true line *249would represent the true acreage, and the fictitious line would include a per cent, of fictitious acreage. Hence, where the acreage to be ascertained is calculated on paper upon the basis of required distances, and the measurements are all upon mountain surfaces, the method in question produces a per cent, of error and loss of land. We therefore think the method involved in the G-ile survey was not the proper method for ascertaining the boundary and divisional line between the parties and the land conveyed, and upon all the evidence we find as a matter of fact that the Gile survey does not set off the 10,000 acres which the deed calls for. The testimony satisfies us that long observation and experience among practical surveyors in the locality in question has demonstrated that some reasonable allowance should be made, and ordinarily and customarily is made, for losses which inevitably and necessarily result in measuring land with a tape or chain in thickly-wooded, rough, and precipitous places. This is not to get more acreage than the deed or grant calls for, but to prevent undermeasurement, and to secure the nearest practical and approximate accuracy and exactness. Any other method would disturb and overthrow results and transactions based upon the experience and observations of many years. This is a method or rule of necessity established by experience. The allowances made necessary by the offsets to get around obstructions and by the impossibility of accurately leveling the chain or tape over uneven and precipitous territory are not gratuitous allowances of land, but compensatory allowances of land, supposed to be equal to the estimated error involved in the erroneously elongated line of measurement,—a somewhat inexact and uncertain process, but one involving a necessary element of discretion as an inherent part of the survey and measurement contemplated. The method is not adopted to get more land than the parties intended, or less than the parties intended, but to get the nearest practical approach to exactness and to the acreage called for by the deed. From the nature of the work, there can be no such practical result as absolute exactness. Át least an attempt at absolute exactness would present a problem in engineering and surveying more difficult to solve than the parties intended. A more exact result might, perhaps, be reached by triangulation and the use of theodolitic or other nice processes known in engineering and surveying, but this would involve a degree of nicety not ordinarily intended in contemplated surveys of wild lands in the locality in question. The instrument commonly employed is the ordinary compass and chain, not the theodolite; and the method involves an ordinary compass line with careful measurements and reasonable allowance for losses, and this is the practical method intended by the parties, and for which the deed, under the circumstances of this case, calls. The same surveyor with this method, however, cannot measure a line nine or ten miles in length through a rough, wild country, and, removing all marks of such survey and measurement, make a second survey and measurement, reaching precisely the same result. From the necessities of the situation and undertaking, therefore, practical and reasonable exactness is the only result that can be expected. Of course, it goes without saying that, where absolute exactness is possible and reasonable under *250the circumstances, such result is the one called for by the deed or grant; but, when this cannot be had, and was not expected, then the nearest practical approach to it is the intended and true result. Holding these views, we must find that the Gile survey, while unquestionably made with care and efforts at exactness under the method adopted, incloses less land than the deed calls for, and less than the parties intended.
In conclusion we find that neither party has established a divisional line, and that neither party is entitled to a verdict upon the evidence as it now stands.
Upon the foregoing findings of fact, we hold that the deed does not convey any land north of a line extended on the course of the southerly line of Irving’s location. We also hold, upon the findings, that the deed operates as a conveyance of land between a line to be extended on the course of the south line of Irving’s location and the lines of Dummer and Millsfield and Stark, and that the land conveyed extends far enough westerly to include 10,000 acres; and that the necessary line from the southwest corner of Irving’s location to intersect a necessary line drawn from Stark parallel with the east line of Dummer and Millsfield, may and should be supplied in order to allow the deed to operate as the parties intended. Land Co. v. Saunders, 103 U. S. 316, 322; Winnipisiogee Paper Co. v. New Hampshire Land Co., 59 Fed. 542, 547.
The Glen Manufacturing Company, in this trial, stands upon a supposed agreed line, rather than upon proof of a true line marking off the exact 10,000 acres in accordance with the deed; and, without determining the question whether it includes more or less than 10,000 acres, we hold that it is not an agreed line, or a completed survey under the deed, and therefore does not establish the extent of the grant or deed. Bartlett v. Young, 63 H. H. 265, which is the latest expression of the supreme court of Hew Hampshire on the subject of agreed real-estate lines, fully recognizes the idea that the agreement must be executed. This, as we understand it, means more than that the parol agreement must be executed so far as words are concerned. It means that there shall be performance under the agreement, if performance is contemplated. If the agreement is to run an agreed line, the line must be run. If there are no monuments, and the agreement is to establish corner-stone monuments, that must be done. In other words, the act or acts contemplated—the thing agreed to—must be executed. That is what is meant by executed parol agreements binding upon parties and their descendants in title to real estate. A rule of departure from this requirement would be fraught with more uncertainty, greater insecurity, and greater danger than could be foretold. The element that removes the agreement from the operative effect of the statute of frauds is not that the word-agreement is executed, but that the act affixing the agreed boundary to the land itself is executed. In one case it would be a naked parol agreement, and in the other a parol agreement, strengthened and executed by connecting it with the land by the erection of monuments which become a part of the land as physical and visible *251boundaries. Placing a line upon a plan is not a compliance with the requirements of this rule. The rule that a deed is certain which can be made certain (Corbett v. Norcross, 35 N. H. 99; Wells v. Iron Co., 47 N. H. 235, 239; Land Co. v. Tilton, 19 Fed. 73, 77), and that unsurveyed lands may be conveyed by a paper-plan allotment, is based upon the idea that the plan by express reference becomes a part of the deed itself. It is quite another thing where a party relies upon •a plan not referred to by the deed, and not signed by the parties. The doctrine that a deed is certain which can be made certain does not apply to a plan not referred to in the deed, and one which is founded in naked parol; and the doctrine of executed agreed lines does not mean that agreements are executed which can be executed. The agreement, so far as it related to 'a line upon the ground from the southwest •corner of Irving’s location to Williams’ west line, was wholly unexecuted, and to that extent rests in naked parol; and therefore, in any future controversy which should develop no line or boundary upon •the ground, the only possible answer would be that the parties agreed a line should be run ; and this answer that the agreement was to be -executed would not bring it within the requirements of the law which gives force to executed parol agreements establishing real-estate boundaries.
The Weston Lumber Company stands upon an ex parte line which they say is a true line, and in respect to this we hold, upon the foregoing findings of fact—First, that the line is not the bound contemplat•ed by the conveyance. This ruling is based upon the idea that the .line is not laid upon the ground in conformity with the deed in this: that in going north of a line extended on the course of the south line ■of Irving’s location, and to 'Columbia, it includes territory not intend■ed to be conveyed.
And secondly, if the line extendingnortherly to Columbia line should, in any event, be held to answer the calls of the deed, then the question would come as to the area of land, and whether the line incloses the number of acres conveyed. Surveys and measurements, as applied to this question, are not absolute, but are competent evidence bearing upon the controverted question of quantity. The problem of ascertaining an unknown and unlocated exterior bound of a definite ■quantity of land when three of the exterior bounds, but no distances, are given, must be reasonably solved. In such a case, where the '.method adopted for the solution of the problem involves measure-•ments of the given lines, and a survey and measure of the line to be ascertained and located, the question whether the method adopted was reasonable, and the question whether the particular surveys and measurements involved were reasonable, are questions of fact, which •are to be found in the ordinary way upon the evidence in the case. 'The line called for by the deed is a straight, horizontal line, and the •number of acres called for is 10,000, and the means employed for as•certaining this line and its length and the area of land, whether based upon allowances or the method of leveling the chain without allow.ances, must be reasonable, and the question of reasonableness is to be determined as a question of fact. The means of ascertaining the -true horizontal line and its length necessarily involve an element of *252discretion, but the means employed must always be fair, and the exercise of discretion always reasonable, under the circumstances of the particular survey and measurement. Reasonable exactness is, of course, required, in order to allow the deed to operate as the parties originally intended it should. The fact is found that the Gile line does not include 10,000 acres of land, so again, and upon this ground, we hold that this line.is not the boundary line between the parties.
Under the Hew Hampshire rule, that a deed is certain which can be made certain, and that such a deed operates as a conveyance, the Glen Manufacturing Company holds a valid title to 10,000 acres of land limited on the west and north by the lines to which we have referred, and such lines exist in contemplation of law, although, upon the evidence and the findings, the whereabouts are not ascertained. The parties holding this deed are, upon the findings, in actual possession of at least some portion of the territory, and in constructive possession of all that the deed conveys; while the Weston Lumber Company, being in actual possession of some portion of the to.wnship of Odell, is in constructive possession of all that the Thompson deed does not cover. In each of the cases under consideration the burden is upon the plaintiff .to show that the defendant has entered his close, and, failing to establish that his domain includes the territory on which the acts in question were committed, he fails in his proofs, and therefore cannot recover. The parties acted upon the idea (and probably correctly) that the obligation was upon the grantor to make the survey, and locate the 10,000 acres, thereby rendering the deed certain and operative upon a particular part of the township. The grantors of the Weston Lumber Company, having conveyed to Thompson, under whom the Glen Manufacturing Company claim, 10,-000 acres of the township of Odell, have so commingled the interests by creating a joint or interdependent constructive possession that neither party can establish exclusive title to the locus in quo, and therefore neither party is entitled to recover in trespass until the deed is made certain by ascertaining the true location of the contemplated divisional line, and this is for the reason, not that neither has title, or that neither has trespassed, but for the reason that neither shows title, nor exclusive right of possession, to the disputed territory as against the other. In other words, neither supplies the burden of proof by showing the extent of his own territory or the limit of his adversary’s territory. While the deed is upheld as a muniment of title to the 10,000 acres, the title or right of possession to the particular locus in question is left uncertain upon the proofs. Ordinarily, upon such findings and rulings as are here presented, judgment would pass for the defendant in each case, but, in view of the great length and expense of the trial, and of the holding that the divisional line exists as a matter of law, although incapable of ascertainment upon the proofs and findings, we are inclined to defer judgment to the end that the parties may agree upon a surveyor or surveyors to mark the boundary in accordance with the foregoing findings and the construction which we have given to the deed; and, in the event of the inability of the parties to so agree, we will consider a motion from either side directed to the appointment of a surveyor or surveyors for *253that purpose upon consent, and, that failing, to open the case for further and additional evidence as to the whereabouts of the true divisional line, including, perhaps, evidence of a survey upon notice to the adverse party, to be considered in connection with the evidence already before us.