delivered the opinion of the court.
Designating the parties as they stood in the trial court, plaintiffs and defendant are coterminous owners of lands in King William county.
This is a statutory proceeding, Code, section 5490, brought to ascertain and fix their division line. That line originally ran with Manquin creek, and still runs with it unless the course of the creek itself has changed.
Plaintiffs hold by inheritance from their father, J. M. Abrams, Sr., who took title by deed from R. V. Florence, special commissioner, appointed in the cause of Ida B. Blake et al. v. Mary E. Lincoln et al., lately pending in the Circuit Court of King William county. The deed itself bears date June 17, 1920, and contains no detailed description of the land conveyed. It had, however, been in the Blake family for very many years. The records in King William county were burned in 1885, but in the processioners’ book for that year it is described as bounded in part by Manquin creek.
Thomas Carter, who owned Pampatike plantation, died about 1885. His estate was administered in the chancery cause of C. Shirley Carter et al. v. Ann W. Carter, widow of Thomas Carter, et al. By decree of date October 5, 1885, Roger Gregory and others were appointed commissioners to make partition. This they did. Their report was con*687firmed by decree of date April 19, 1886. Lot “A,” containing 478% acres, fell to Mary Nelson Carter Buckner. Its line runs “to the mouth of Manquin creek thence up said creek 226 chains to a gum tree corner Fountainbleau.” Under conditions not necessary to detail it passed to Evelyn Byrd Buckner Bassinger and others, who by deed of date January 10, 1930, conveyed the same to C. C. Woody, defendant here. The lines follow those set out in said decree of April 19, 1886. There is no doubt whatever that the division line between the Carter and the Blake lands ran with Manquin creek. They were old estates and that situation had remained unchanged for probably more than a century. Plaintiffs contend, however, that the course of this creek itself changed or was changed about 1904, or within a comparatively short time thereafter, and that it now flows, not in its ancient bed, but in a new channel, which, if accepted as the true line, takes from them about twenty-eight and one-half acres of land and adds it to Woody’s holdings. This Woody denies. He contends that the creek is now where it has been for an indefinite time.
Manquin creek flows in a southerly direction. On its left bank is defendant’s land, on its right the plaintiffs’, ‘who claim that on the Pampatike side stood a dike built along its contour and thrown up to protect that estate from flood waters which ran across it after heavy rains. They further claim that the old creek bed was filled up by sawdust from a lumber mill two miles away, in operation from 1901 to 1904, or by the cutting of a ditch that diverted water from the old channel into a spring branch flowing west of and somewhat parallel to the old creek and into Blake’s pond, which stream is now the main channel of that creek.
L. D. Robinson, at the request of the plaintiffs, and under authority of Code, section 5490, has made a plat of the land in controversy, which plat, in a general way, shows the claims of the parties. From that plat it appears that the dike follows the old creek and he tells us that this physical fact is patent on the ground.
*688J. Hooper Edwards, born in 1864, had lived in this neighborhood and had known these properties all of his life. He said that the dike ran along the old creek on its Pampatike side; that this old creek bed is now filled up, and that he first noticed evidence of change when hunting over the land with Gena Clements. “He called his attention to it, and he said ‘Yes the old Manquin creek is gone for good,’ a part of it was running down the black land (Pampatike) and the rest was going around into Max Spring swamp.” Clements went to Pampatike in 1902 but Edwards did not recall how long thereafter it was that he noticed this change.
Dr. William Croxton, now of Richmond, was born in King William county. As a boy, in 1891 or 1892, he was hunting with his father on the Black Pampatike land and held his father’s horse. He tells us that “the dike was up by the creek, immediately below it.”
George Moore worked at the sawmill. It was about two miles up the creek. He remembers that sawdust was dumped into it until some one objected.
W. B. Atkinson in 1916 cut ash timber on this land, bought (from Mr. Blake. He saw evidence of an old creek bed and ■ of a dike by it. How old it was he does not undertake to : say. .The creek itself runs now as it did when he saw it.
Hilliard Temple, sixty-one years old, has known these ■ properties all his life. He said that the creek has changed . in the past twenty-five years, and that no water now runs in the' old bed which ran with the dike. He hauled lumber ■from the sawmill to the Pamunkey river. A tramway ran . down with the creek and across it four or five times—that is, across the old creek in which water was then still running. He also tells of sawdust being thrown into the river until that, custom was stopped by Col. Carter. Five or six years later, when-hunting on this land, he noticed that water was ■ breaking through the creek and running over on Pampatike. On pne occasion hje killed a turkey on this lowground, which was- then known as “Blake’s Meadow.” The substance of this witness’ evidence is that from 1901 until 1904 this *689creek ran in a channel through which it no longer flows. He said that “Mr. Woody took me through to the creek where it is now and said, ‘You know this is the old creek.’ I said, ‘I have been through here hundreds of times but never saw a creek here before.’ That is the first time I seen a creek across there.”
William Allen is sixty-one years old and lived on Pampatike for twenty-two years. He went there thirty-eight years ago. He said that at that time what is known as Manquin creek ran by the old dike but that it no longer runs in the old channel. When he first knew it, it did not run into Blake’s pond at all. He tells us about the sawdust being thrown into it, and he tells us also that the course of the creek was changed by a ditch which turned the water from Pampatike towards the Max Spring branch. It was cut by Mr. Clements. “He cut that ditch to turn it off the other way towards Blake’s land or across Blake’s land,” and that it was this which changed the course of the creek.
J. R. Adams lived on Pampatike for about eighteen years and left there about 1907. He said that Mr. Clements cut this ditch or canal while he was there and changed the course of the creek. He further said that a dike generally followed the bed of the old creek to protect the black lands of Pampatike.
Ephriam Bowler, eighty years old, testified at the first trial. He was dead when the case was again heard but this evidence was. read to the jury. He was living on Pampatike when the slaves were freed, and left there during the year of the Johnstown flood. He said “that he has worked on this dike many a day and many a year for Col. Carter, and the creek which ran down beside this dike was called Man-quin creek, and that the dike was built to keep the water out of the black land on Pampatike'; that the old creek is now dry, but that there used to be water in it nine or ten feet deep in some places; that Col. Carter called the land between the bed of the' creek and the Blake’s pond, ‘Blake’s Meadow;’ that Col. Carter told him that half of the creek belonged to *690the Blakes and half to Col. Carter; that the old creek never did run into Blake’s pond except during the high water, and the channel of the old creek is the true boundary line between Blake’s and Pampatike; that the dike is close to this creek; that he could plow to end of row and look into the creek; that he moved from Pampatike in 1889, but rented some of the land after Mr. C. C. Woody’s father left Pampatike for several years, and that water was running down the old creek at that time; that the old creek has changed since he stopped working on Pampatike farm.”
T. M. Abrams sunk pits in the bed of the old creek. On one occasion he found a plank, on another a part of an old musket. His evidence is that from sand and deposit it clearly appears to have been the bed of the old stream.
Roger Gregory, sixty-seven years old, has known these properties for fifty-two years. He went to school at Pampatike. His evidence is that the dike ran along its black lands, that the creek ran along the dike, and that Col. Carter had the bed of this creek cleaned out each year to prevent the accumulations of sand and debris. The evidence of other witnesses for the plaintiffs is to the same general effect as that which has been heretofore noted for the defendant, A. D. Willis, who has known this property for fifty years, tells us that the course of the creek has not changed.
Charles Reed rented land on Pampatike in 1896. He says that the creek runs now as it ran then. “I don’t think it has moved an inch.”
W. E. Rougie ran a sawmill on Pampatike in 1896 and 1897, and farmed there from 1904 to 1907. He tells us that the creek has not changed its bed.
W. B. Pointer, sixty-nine years old, has lived in King William all his life and knows this land well, having hunted on it from the time he was eighteen or twenty years old. He said that this creek runs where it ran when he first knew it unless it has changed in the last four years.
Charles Saunders says that he has known this creek well *691for thirty-eight or forty years and that it.is now where it always was.
E. W. Farmer tells us that he has known it for thirty years and hunted along it, and it is where it always was.
J. L. Adams went to live on the Pampatike farm in 1901 and has hunted and trapped this creek for many a day. He tells us that it is in the same old place.
Charles Gaines was one of the men who helped cut the ditch of which we have been told. He said that it did not change the course of the creek. Jim Hines, sixty-nine years old, lived on Pampatike thirty or forty years ago, and was raised on it. He tells us that Manquin creek is there now as it was then.
W. V. Fox has known this creek since 1905. His evidence is that it has not changed its channel in his recollection.
Scott Banks helped cut the Clements ditch. He said it merely took an elbow out of the creek and that the creek itself now runs as it did when it was cut.
E. S. Carter says that the creek along the line in controversy runs as it did thirty years ago.
It is not easy to see why sawdust from a county sawmill two miles away, put into a creek, would change a channel at this particular place. It may have been changed by the ditch from it to the Mac Spring swamp, and it may have been from that tendency of the creek to fill up, noted in the evidence of Mr. Roger Gregory. To prevent clogging, Col. Carter had to clear the channel every year. But we are not primarily concerned with causes. If the creek did in fact fill up, the reasons therefor are not important. Witness after witness whose evidence is not impeached, tell us that there has, in fact, been a change. A number of other witnesses, whose opportunities of observation are equally good, tell us that there has been none. In such circumstances the court has no choice but to follow the jury’s verdict, confirmed, as we have seen, by the trial court. Moreover, the jurors went upon the premises.
What effect, if any, does change in the channel of a *692creek have upon the title of riparian owners? Erosive accretions attach to the land on which they fall. Accessions or abstractions, particle on particle, are so necessary and so general that title cannot be made to depend upon them. When this is all, a stream once a line continues to be the line. But where there is avulsion or sudden change from any cause, natural or artificial, by which a stream leaves its old bed and cuts for itself a new channel, the rule is otherwise, for title cannot be made to depend upon the meanderings of vagrant streams.
In Arkansas v. Tennessee, 246 U. S. 158, 38 S. Ct. 301, 304, 62 L. Ed. 638, L. R. A. 1918D, 258, the court said: “* * * when the bed and channel are changed by the natural and gradual processes known as erosion and accretion, the boundary line follows the varying course of the stream; while if the stream from any cause, natural or artificial, suddenly leaves its old bed and forms a new one by the process known as an avulsion, the resulting change of channel works no change of boundary, which remains in the middle of the old channel, although no water may be flowing in it, and irrespective of subsequent changes in the new channel.”
This statement of the law appears in Nebraska v. Iowa, 143 U. S. 359, 12 S. Ct. 396, 398, 36 L. Ed. 186, approved and copied from Vattel (1 Vattel, Droit des Gens, ch. 22, sections 268-270) : “But if, instead of a gradual and progressive change of its bed, the river, by an accident merely natural, turns entirely out of its course and runs into one of the two neighboring States, the bed which it has abandoned becomes thenceforward their boundary, and remains the property of the former owner of the river, the river itself is, as it were, annihilated in all that part while it is reproduced in its new bed, and there belongs only to the State in which it flows.”
The court went on to say: “The result of these authorities puts it beyond doubt that accretion on an ordinary river would leave the boundary between two States the vary*693ing center of the channel, and that avulsion would establish a fixed boundary, to wit: the center of the abandoned channel.”
These rules apply whether we are dealing with public or with private interests. Oklahoma v. Texas, 268 U. S. 252, 45 S. Ct. 497, 69 L. Ed. 937.
“States and individuals alike are subject to the losses and gains of erosion and accretions; but neither can have the boundaries of his domain changed by avulsion, or by the diversion of the water effected by human agencies.” State of Wisconsin v. Bowen, 149 Wis. 203, 135 N. W. 494, 496, 39 L. R. A. (N. S.) 200. See also, Mitchell V. Baratta, 17 Gratt. (58 Va.) 445.
It is perfectly clear that such change as has been brought about was not due to erosion.
Error is assigned because the plat of September 10, 1930, made by Robinson, surveyor, was received in evidence.
Code, section 5490, provides that “A plat showing such real estate and boundary line or lines, filed with the petition, may serve the purposes of such description.” It may be conceded that this plat does not tell all we would like to know, but it does tell us what the plaintiffs claimed, and is supported by evidence which they introduced. Even if we should be of opinion that it was highly inaccurate it would still be evidence to be received for what it was worth.
Objection is made to certain evidence dealing with the locations of monuments and the claims of former owners, because it is hearsay.
J. M. Abrams and Mordecai Abrams testified that Bob Blake and Gene Clements, men dead when this testimony was given, told, them where the bed of the old creek lay, and showed it to them on the ground. Ephraim Bowler, who had himself worked upon the dike, testified to the line pointed out to him by Col. Carter. On the same general subject, Dr. Croxton told of statements made by his father, the owner of an adjoining tract of land. There is other testimony of the same general character.
*694These statements were made by men now dead, by former owners of these properties, by their tenants and by neighbors who were particularly qualified to speak on the subject. They are admissible, and, although hearsay, evidence of this character is often the only evidence it is possible to obtain. It is a rule of necessity and gives to us the best evidence available. Monuments perish and streams fill up. At times if they cannot thus be relocated they cannot be relocated at all.
“That boundaries may be proved by hearsay testimony is a rule well settled; and the necessity or propriety of which is not now questioned. Some difference of opinion may exist as to the application of this rule, but there can be none as to its legal force.” Boardman v. Reed, 6 Pet. 328, 341, 8 L. Ed. 415.
In Harriman v. Brown, 8 Leigh (35 Va.) 697, Judge Tucker said: “I hold then, clearly, that evidence is admissible to prove the declaration of a deceased person as to the identity of a particular corner tree or boundary, provided such person had peculiar means of knowing the fact; as, for instance, the surveyor or chain carrier who were engaged upon the original survey, or the owner of the tract, or of an adjoining tract calling for the same boundaries; and so of tenants, processioners, and others whose duty or interest would lead them to diligent enquiry and accurate information of the fact; always, however, excluding those declarations which are liable to the suspicion of bias from interest.” See also Clements v. Kyles, 13 Gratt. (54 Va.) 468; Hurley v. Shortridge, 118 Va. 136, 86 S. E. 858; Annotations, Harriman v. Brown, supra, Va. Rep. Anno.; Wigmore on Ev., sections 1563-4-5. .
There was introduced a plat of Pampatike farm, made by L. D. Robinson in 1885 when commissioners were appointed to make partition of that estate. Mr. S. S. Robinson, the son of the surveyor and himself a surveyor, was asked to explain what certain lines on that plat meant. He might well have been permitted to answer, for in cases of *695this character courts should be liberal in their reception of evidence.
Much was said about the course of the creek, as shown on the 1885 plat. On that point the plat itself is silent. The substance of the evidence rejected is that Mr. Robinson was not permitted to testify as to these compass directions. The court in conclusion said: “The map positively does not show whether it (course of the creek) is southwest or not and you have all that evidence before the jury by witnesses who have testified here. You have examined them as to whether that creek flows south or southwest. You have that evidence definitely from witnesses who have testified. If Mr. Robinson can say, I can take that plat and that plat positively indicates the direction in which it flows, I would let it go to the jury; otherwise I don’t think it ought to go to the jury. The plat speaks for itself.” In other words, the court was unwilling to let this witness state what he thought the probabilities were. Wide latitude had been theretofore extended to other witnesses and might have been extended to him; but this, if error at all, is not reversible error. No record is perfect. This case has been twice tried and should not be retried unless justice plainly demands it. Oliver v. Commonwealth, 151 Va. 533, 541, 145 S. E. 307.
Instruction B, tendered on behalf of the defendant and refused, is predicated upon the idea that the change, if any, was brought about by the imperceptible processes of erosion. There was no evidence to sustain it, and it was properly rejected. There is evidence to show that this change was either due to the dumping of sawdust, to the digging of an artificial channel or to the deposit of silt. It is not necessary for us to know which one of these causes the jury thought had controlled. They did think that the old channel was abandoned and a new one carved out.
This is rejected instruction C: “The court instructs the jury that if they believe from the evidence that the said defendant and those under whom he claims to have been in actual, open, exclusive and hostile possession of the said *696land in dispute for fifteen years, next preceding the commencement of this suit, under claim of title, then the said defendant is entitled to the said land and the jury must find for the defendant.”
Plaintiffs did not hold under color of title, for under the partition of 1886 the Buckner land extended to Manquin creek and all subsequent links in title have adopted the boundaries there set out. We have theretofore seen just what was meant by Manquin creek and color went only to it.
If claim of title is relied upon, it must have been hostile and notorious. Moreover, there must have been actual and not constructive possession. A part of this land was enclosed. Mr. Woody testified that its fence was put there by Mr. Clements that he might graze his cattle. Woody said that this was all that he knew about it. T. M. Abrams and Mordecai Abrams say that this enclosure was built by Mr. Clements under permission—that is to say, they testify that there was nothing hostile about it, and that since it was not hostile title cannot be traced to it.
Instruction D was also rejected. It reads: “The court instructs the jury that the question before them is the location of the true boundary line between the land of the plaintiffs and the land of the defendant, and they are to determine that from all of the evidence in the case, and if from all of the evidence in the case they believe that Man-quin creek is the true boundary line, then they should fix said creek as the boundary line between the said land.”
This instruction was confusing and should have been rejected. When Manquin creek was spoken of as the true boundary line, the jury might have thought that the court was speaking about the creek as it ran in 1886, and they might have thought that the creek as it runs to-day was in mind.
Instructions 1 and 4, given on behalf of the plaintiffs, are objected to.
Instruction 4 deals with adversary possession and tells *697the jury that there was no evidence in the case on which a verdict could properly rest. There was none.
Instruction 1 was proper under plaintiffs’ theory of their case and under the law as we have stated it.
We find no reversible error in the judgment of the trial court and it is accordingly affirmed.
Affirmed.