reviewing the facts found to be established, delivered the opinion of the court.
This case was originally decided adversely to the plaintiff on findings of fact and conclusion of law, with a memorandum, but without opinion. Plaintiff has moved for certain requested amendments to the findings of fact and, conditionally, for a new trial. In response to the plaintiff’s motion, sustained in part only, new findings of fact have *123been made for the purpose of removing any possible inaccuracies, and, of its own motion, the court has made the findings more in detail in some respects to the end that, as the court sees it, they may present as accurate and detailed picture of the situation with which we have to deal as a record, unsatisfactory in many respects, will permit.
The plaintiff brought its action in this court on the 5th of May, 1911, alleging the taking of parts of a number of roads along or near to the Monongahela River in Marion County, W. Va., by reason of the construction of Dams 14 and 15 by the defendant in aid of navigation, under authority of Congress, and sought recovery under the Fifth Amendment to the Constitution.
The two dams were built about the same time. It is shown that Dam 15 was completed and the pool began to fill October 6, 1903, and the same condition prevailed as to Dam 14 on November 3, 1903. Dam 15 was located a short distance below the city of Fairmont and about 2 miles below the Fairmont River gauge. Dam 14 was below Dam 15, the distance not being shown, but it was at such distance that it was below the roads with which we now have -to do and within such distance that the “pool” created thereby was along these roads. Pool 14 was necessarily below Dam 15 and between the two dams.
Certain roads declared on in the petition were permanently flooded at pool level when the dams were put iri operation, and as to these, described in the findings, claim was abandoned upon trial of the case because of the statute of limitations.
The roads remaining for consideration, since they are connecting, constitute practically one continuous road, or, more properly speaking, a part of one continuous road, the part involved being in the aggregate about 2f miles in length and extending from the described beginning point at the upper end to Parkers Run, although the upper 2 miles terminating at Pharaohs Run, the lower corporation line of Rivesville, is referred to as a turnpike road, while the lower three-quarters of a mile from Pharaohs Run to Parkers Run is referred to as a county road. Both must be assumed to be in fact county roads, and the rather persistent use in the *124record of the word “turnpike” with reference to one of them justifies the suggestion that, so far as the usually accepted significance of the word is concerned, it is a misnomer, finding justification only in its derivation from the original name. The term “ turnpike ” in its original and strict sense had no meaning whatever indicative of the character of the roadway. It referred to the gate maintained for the collection of tolls, the “ pike ” or pole which was turned. Since roads upon which tolls were collected came usually to be better improved than other roads, generally macadamized or graveled, the word came to be regarded, in its usual significance, as importing something with reference to the character of the road, and it is in that sense that it is a misnomer as to this road. The findings with full degree of liberality show that “ except where macadamized for a short distance ” it was a dirt road. It was not necessary to show that the “ macadamizing ” for that short distance was an accident of conditions encountered rather than an intent to improve and was of such a character that travel sought to avoid it by the use of a byroad.
The portion of road declared upon and now in question began where the road leading down from the city of Fair-mont and running a considerable distance back from the river, but gradually approaching it, crossed the railroad from the landward side to the river side thereof, and began to follow the course of the river. The Monongahela Fiver follows a tortuous course. At this beginning point it is flowing north, shortly below bends slightly to the left, and then gradually bends to the right, describing approximately an arc of a circle, and when opposite the lower section of road with which we have to do it flows slightly south of east, but, taking directions from the beginning point, the roads are' described as on the west bank of the river. The findings give a fairly accurate idea of the contour of these roads with their ascending and descending grades, consequent varying elevations, and the elevations of low places in comparison with pool level opposite.
These connecting roads were entirely along and opposite pool 14 created by Dam 14 and evidently wholly unaffected by the water in pool 15 at poool level. Dam 15 was a short *125distance, not accurately shown, above the beginning point of the portion of these roads here involved, and what effect it had on the level of the waters below when the river was above pool stage is not directly shown.
It will be noticed that some doubt is cast, in the findings, on the accuracy of the statement as to the pool level of pool 14. It is taken as corresponding to 14.3 feet on the Fair-mont gauge. The testimony with reference to the matter, the fact that elevations above sea level show pool 14 to be 10.7 feet below pool 15, and other circumstances, cast grave doubt on the matter. And in this connection it may not be inappropriate to suggest that a record so deficient and confused in important respects as to cast upon a court such a burden in acquiring a satisfactory conception of a situation as has been done in this case is inexcusable. But so far as this feature of the matter is concerned it seems that the plaintiff’s engineer witness who submitted profiles which are in evidence for what they are worth and testified as to the elevation of the low places in the roads used those figures as the basis of his testimony, and a comparison of road elevations with pool level may therefore be accurate even though the basis be in point of fact in error. Here it may be said, as it could not properly be said in the findings, that the elevations of the low places are stated in the findings at 1.2 feet above the elevation given by this witness in his testimony because of his afteradmission that his profiles, from which he was testifying, were erroneous to that extent in the indicated elevation of his base line, an error which he attempts to, but does not properly, correct in his profiles subsequently filed for that purpose.
These roads, it is shown, were subject to occasional overflows at places before these dams were built, interfering temporarily with travel, and once, a few years' before, they had been entirely submerged. They were along the Monongahela Biver, whose characteristics as a swiftly flowing river in a mountainous section subject to rapid rises and falls and the natural effects thereof are matters which the court may know. After the construction of the dam rises to given stages were more frequent, high stages were somewhat more prolonged, although they, for the most part, were at*126tained and subsided within the day, and coincident, at least in point of time, therewith the deterioration of these roads by washing's and cavings of the banks became more rapid. Banks in course of time caved at places to and, at other places, into the roadway and it is not found that they had theretofore caved to such an extent as to injure the road. It is not shown, in fact, that the banks along the old turnpike had caved before the building of these dams, but it is shown and found that trees growing along the bank had caved off, a distinction to conform to the testimony without a substantial difference. It is not shown that the banks along the county road had caved to the extent of encroaching on or injuring the road before the dams were built, but it is shown that the banks were unprotected and were of a sandy clay easily washed, and the court is not required to indulge a presumption in conflict with common knowledge as to what must result under such circumstances. Banks so situated along such rivers always have washed and always will wash. It is the natural result of the action of the water which the court must know and of which it is not required to assume ignorance and it is especially true where the bank in question, as in this case, is on the concave side of a pronounced curve.
We are furnished with evidence tending to show the condition of these roads in 1911, 1912, and 1915, when the testimony was taken, and are asked to conclude that their deterioration after 1903 and alleged final destruction was due to these dams. We are not furnished with any evidence except inferential as to their condition in 1903 compared with corresponding periods before that year. The defendants did not see fit to take testimony.
In this connection it is proper to observe that the plaintiff, in its petition filed May 5,1911, then alleged a “ taking ” of the roads in question. A part of the allegation, applicable to these roads not permanently submerged, is as follows:
“ and that the rises in the waters of said river, since the completion of said locks and dams, have occurred, and are now occurring, at such frequent intervals and for such duration as to render said roads, bridges, and public landings valueless to said county and to entirely destroy and wash them away, taking from said county the use, occupation, and *127enjoyment thereof; and that there is now such a serious and continuous interruption to the common and necessary use and enjoyment of that part of said property not permanently submerged and overflowed, but affected by frequent and continuous pverflows, as amounts to a taking thereof by the United States for public use, and the use thereof changed, by reason whereof an implied promise and contract to pay therefor has accrued in favor of said county and against the United States under which the Federal Government is obligated to pay said county the value thereof.”
It is also averred in the petition—
“ That said county court has used reasonable care and diligence to protect and preserve said roads, bridges, and public landings from such destructive overflows, but has been unable to do so.”
A considerable part of the testimony in this case was taken by the plaintiff in June, 1911, shortly after the commencement of the action. Further testimony was taken by the plaintiff in January, 1912, and still further testimony was taken in October, 1915. During the examination of a witness for the plaintiff on October 25, 1915, counsel for the United States entered of record a motion to strike out all the testimony “ on the ground that the court has not jurisdiction to entertain the same, the taking, if any, having occurred more than six years prior to the filing of the petition in this case,” following which the record shows:
“In reply to which counsel for the county says that the testimony heretofore taken in this case has shown that this pike was in use and, therefore, not taken, but at this hearing it is shown that it has been taken and abandoned since the last taking of testimony.”
In plaintiff’s request for findings of fact, not filed until February 24, 1917, the court is asked to find that—
“From 1903 until 1915 the injuries to said turnpike increased and grew worse each year by the more frequent and destructive overflows. Finally, in 1915, much of said turnpike became impassable and destroyed as the result of the Government work and had to be abandoned.”
But in a request filed May 9, 1917, we are asked to. amend that finding and say—
“The taking of said turnpike did not occur until 1911. During the early part of that year the greater part of said *128turnpike became impassable and valueless to tbe county and had to be abandoned.”
Considering the facts found we are necessarily interested in the further conclusion to be drawn therefrom as to cause and effect. It is found that—
“After the completion of the dams flood waters reached somewhat increased heights, flood heights continued for somewhat longer periods, given stages were reached oftener, and the flood waters of the river overflowed parts of the roadways in question more frequently than theretofore.”
It is not found as a fact that these things were caused by the dams, but in this connection there is a finding as to the number of times the river had reached given stages, as shown by the Fairmont gauge, during several years before and after the construction of the dams, with other pertinent facts as to duration of flood stages. It is also found that—
“ The stages of water below Dana 15 in pool 14 during the high-water periods shown in the above tabulation subsequent to the construction of the dams and filling of the pools is not separately shown, and the court is not infomied as to the effect of Dam 15 on the stage of water below it in pool 14, or the comparative stage of the water in the last-named pool.”
It is our duty to find “ the facts in the case established by the evidence.” That these dams caused the conditions stated, if in fact they did cause them, may be an ultimate fact, but it seems to us rather a conclusion to be drawn, if justified, from the facts proven. But however that may be as an abstract question, we do not find satisfactory evidence justifying this finding therefrom as a proven fact and, if to be considered as established, it must be because it is the required conclusion from the facts which are proven. Those facts are found for the most part in the tabulated statement as to flood heights and some additional facts immediately following, supplemented by some testimony as to increased frequency. We are not unmindful of the fact that witnesses have testified in a general way as to cause and effect as between these dams and existing conditions, but it is for the court to determine the weight to be given the testimony of any witness and expressions of opinion from interested witnesses who are not experts or who do not show that their *129opinions are predicated on all the facts material to be considered are not to be given great weight.
If we were dealing with one dam and with roads lying above that dam or with roads lying above the upper of these two dams, the situation might not be so difficult; but we are not concerned with the effect above the first of these obstructions to the flow of water, but with resultant conditions below that dam, since the roads in question are below it, and the effect, so placed, of the lower of the two dams. What that effect was as to the comparative height of the waters we are not satisfactorily informed and are not shown the way to a satisfactory conclusion based either on facts shown or matters of common knowledge. The statistical data furnished by the plaintiff to show increased frequency in a stated height of the water is data obtained from the Fairmont gauge, about 2 miles above the upper of these dams. It is used to show that a stage of 18 feet was reached much oftener during some years after 1903 than theretofore. Incidentally, it may be said that the directing of testimony, particularly to an 18-foot stage, seems somewhat peculiar, since in no place in plaintiff’s testimony does it appear that such a stage reached the roads in question, and for that reason we have incorporated in the findings, from the same statistics, the facts which appear therein as to higher stages; but, with that condition in mind as to the effect of that stage, the matter is not material to the discussion one way or the other.
With a rainfall in the Monongahela Valley above the Fair-mont gauge which would have produced an 18-foot stage on that gauge before the dams were built, it is probably to be concluded that with the same amount of rainfall, other conditions being the same, the effect of Dam 15, when built, would be to obstruct the flow of the water, even when' above pool level, to such an extent as to indicate, in some degree at least, a higher stage on that gauge. It follows that a more frequent showing of an 18-foot stage on that gauge might be due to either one of two causes. If we assume, for illustrative purposes, that Dam 15 raised the level of the water resulting from a given rainfall 1 foot on the Fairmont gauge over the stage which would have been shown before the oon-*130struction of tlie dam, then a 17-foot stage occurring as often before the dam was built as an 18-foot stage occurred after would, by reason of the effect of the dam, be converted into the more frequently occurring 18-foot stage. If, then, increased frequency was due to this cause, we are in the dark and can not venture to conclude as to the relative conditions below Dam 15, where the roads in. question were located. The other cause of increased frequency might, of course, have been increased frequency in the i’equisite rainfall or changed conditions along watersheds, very commonly discussed as causes contributing to increased heights of flood waters in rivers. These causes, if the real causes, and manifesting themselves below Dam 15 as well as above, are causes for which no responsibility can attach to the defendant.
Upon the question of frequency alone of given flood stages, a term we use with reference to all these high stages of the river, although strictly speaking 25 feet was “ flood stage,” we have not found that increased frequency was due to the dams in question. The facts found do show that given stages, particularly the 18-foot stage about which plaintiff’s witness testifies, occurred oftener after 1903 than during and before that year, but we have declined to conclude that because conditions changed somewhat after the doing of a given act the change was necessarily due to that act. The data itself leaves room for the assumption that other conditions than the dams must have had to do with the results. In 1914 the stage of 18 feet, not very important for consideration, was reached four times as against twice in 1903, but more important stages, 21 feet and 23 feet, were reached the same number of times and peculiarly the high stage reached during each year (23.8 feet) was exactly the same. Two stages of water are given in 1903, one of which passed 18 feet but did not reach 21 feet, while four stages are given in 1904, three of which passed 18 feet but did not reach 21 feet, an increased frequency in the lower stages. In 1905 there were more high waters, each given stage being reached oftener, and the flood stage of 25 feet being exceeded by 4.3 feet, but in 1906 there were fewer of the lower stages, the same number of the intermediate stages, but no 25-foot stage. *131In 1907 there was a marked increase in 18-foot stages, and the 25-foot stage was passed twice, while in 1908 there were only half as many 18-foot stages, one-third as many 23-foot stages, and no 25-foot stage. In 1909 there was an increase by two of 18-foot stages, one-half as many 21-foot stages, and no 23-foot or 25-foot stage. The highest reported stage subsequent to the construction of'the dams was in 1907, the year of the greatest number of high waters, and was 30.7 feet, while a stage of 37 feet is found to have been reached some years before. The average of the heights given for 1904 is slightly less than for 1903, and the average of all heights given after 1903 is less than the average of those before and including that year. It seems, therefore, that when, in a case of this kind, we attempt to go beyond the effect of pool level, the natural and intended results of the construction of the dams, we are necessarily largely in the realm of conjecture. We know there is variation in rainfall from year to year, we know that there are changed conditions along our rivers, the effect of which has been the subject of consideration and discussion by eminent authorities and by those having to do with the resultant problems, and in the light of these things we can not conclude that an increase in a given condition, especially when not continuing in an even •course but subject to marked fluctuations from year to year, is necessarily due to a specific intervening act. For whatever of importance it has it may be added here that the findings do not sustain the allegation of the petition that the plaintiff “ used reasonable care and diligence to protect and preserve said roads,” etc.
But without attempting to rest our conclusion in the case solely upon any question or questions of fact as to the effect of these dams on the waters of the Monongahela River, or upon any failure of the testimony and of the facts found to satisfy us as to the conclusion, in these respects, necessary to be reached to entitle the plaintiff to recover, let us assume, for the sake of the discussion, that these dams were the cause of such increased flood heights and of such increased frequency and duration in flood heights as are shown, and, in the light of the facts as to the effect of the flood *132waters on the roads in question, consider the important legal questions involved and necessary for consideration in determining the plaintiff’s right to recover.
These questions are, first, Do the facts shown amount to a “ taking” under the Constitution? and, second, If so, when did the cause of action accrue?
While we have for our consideration and guidance a number of decisions by the Supreme Court of the United States on the question of a taking under the Constitution, we have no case which undertakes to lay down fixed rules for the determination of the question in all cases, and we have no case which in all’ its features very closely approximates the instant case. It must be our task, therefore, to consider the principles established by the decided cases, harmonize them so far as in us lies, and apply them, as best we may, in the determination of the question as here presented.
In the oft-cited case of Pumpelly v. Green Bay Co., 13 Wall., 166, practically the pioneer decision by our Supreme Court on the question of a taking by overflow without actual conversion to public use, the plaintiff’s land was permanently overflowed by reason of the erection of a dam across the Fox River. The constitution of the State of Wisconsin contained a provision almost identical with the Fifth Amendment to the Constitution of the United States. It was contended that the dam was built by authority of statute, that the damage complained of was such as the State had a right to inflict in the improvement of the navigation of the river, that there was no taking within the meaning of the constitution, and that the damage was a consequential result of such use of a navigable stream as the Government had a right to for the improvement of its navigation.
Speaking for the court, Mr. Justice Miller said:
“ It would be a very curious and unsatisfactory result if, in construing a provision of constitutional law, always understood to have been adopted for protection and security to the rights of the individual as against the Government, and which has received the commendation of jurists, statesmen, and commentators as placing the just principles of the common law on that subject beyond the power of ordinary legislation to change or control them, it shall be held that if the Government refrains from the absolute conversion of real *133property to the uses of the public it can destroy its value entirely, can inflict irreparable and permanent injury to any extent, can, in effect, subject it to total destruction without making any compensation, because, in the narrowest sense of that word, it is not taken for the public use. Such a construction would pervert the constitutional provision into a restriction upon the rights of the citizen, as those rights stood at common law, instead of the Government, and make it an authority for invasion of private right under the pretext of the public good, which had no warrant in the laws or practices of our ancestors.”
After adverting to cases discussing limitations on the right of eminent domain “ as the doctrine was understood before it had the benefit of constitutional sanction” and decisions of State courts holding the doctrine that for consequential injuries to property caused by such public improvements there is no redress, the learned justice said:
“ We are not unaware of the numerous cases in the State courts in which the doctrine has been successfully invoked that for a consequential injury to the property of the individual arising from the prosecution of improvements of roads, streets, rivers, and other highways for the public good there is no redress; and we do not deny that the principle is a sound one, in its proper application, to many injuries to property so originating. And when, in the exercise of our duties here, we shall be called upon to construe other State constitutions we shall not be unmindful of the weight due to the decisions of the courts of those States. But we are of opinion that the decisions referred to have gone to the uttermost limit of sound judicial construction in favor of this principle, and, in some cases, beyond it, and that it remains true that where real estate is actually invaded by superin-duced additions of water, earth, sand, or other material, or by having any artificial structure placed on it, so as to effectually destroy or impair its usefulness, it is a taking, within the meaning of the Constitution, and that this proposition is not in conflict with the weight of judicial authority in this country, and certainly not with sound principle. Beyond this we do not go, and this case calls us to go no further.”
It is thus quite clear that this case lays down the rule that an actual and continuing invasion of one’s property by super-induced additions of water effectually destroying or impairing its usefulness constitutes a taking under the Constitution, though there be no actual conversion to public use.
*134Predicated largely on the Pumpelly case as authority on the question of whether or not there was a taking is the case of Lynah, 188 U. S., 445. The property alleged to have been taken was a rice plantation on the Savannah Biver, lying, in elevation, between mean low and high tides. For the purposes of rice culture it was protected from overflows by an embankment along the river and intersected by canals or waterways with flood gates therein to permit the flooding of the land at high tide, when desired, and its drainage at low tide. In its work of improving navigation the United States built in the Savannah Biver, below this plantation, certain dams, training walls, and other obstructions, the effect of which, it is said in the findings, was—
“ to raise the level of the Savannah River at this plantation, and to keep the point of mean low water above it natural point, so that the outlet of the trunks and waterways above spoken of in the bank of said plantation, instead of being above this point of low-water mark, is now below this point. Another direct result was that by seepage and percolation the water rose in the plantation until the water level in the land gradually rose to the height of the increased watei level in the river, and the superinduced addition of water in the plantation was about 18 inches thereby. By reason of this it gradually became difficult, and has now become impossible, to let off the water on this plantation or to drain the same, so that these acres dedicated to the culture of rice have become boggy, unfit for cultivation and impossible to be cultivated in rice.”
The next finding, over the meaning of which a difference of opinion arises, is as follows:
“By the raising of the level of the Savannah Biver by these dams and obstructions the water thereof has been backed up against the embankment on the river and has been caused to flow back upon and in this plantation above the obstruction, and has actually invaded said plantation, directly raising the water in said plantation about 18 inches, which it is impossible to remove from said plantation. This flooding is the permanent condition now, and the rice plantation is thereby practically destroyed for the purpose of rice culture or any other known agriculture, and is an irreclaimable bog and has no value.”
Interpreting the findings, the opinion of the court says:
“ Some question is made as to the meaning of the findings. It appears from the fifth finding, as amended, that a large *135portion of the land flooded was in its natural condition between high-water mark and low-water mark, and was subject to overflow as the water passed from one stage to the other; that this natural overflow was stopped by an embankment, and in lieu thereof, by means of floodgates, the land was flooded and drained at the will of the owner. From this it is contended that the only result of the raising of the level of the river by the Government works was to take away the possibility of drainage. But findings 9 and 10 show that, both by seepage and percolation through the embankment, and an actual flowing upon the plantation above the obstruction (italics ours), the water has been raised in the plantation about 18 inches, that it is impossible to remove this overflow of water, and, as a consequence, the property has become an irreclaimable bog, unfit for the purpose of rice culture or any other known agriculture, and deprived of all value. It is clear from these findings that what was a valuable rice plantation has been permanently flooded, wholly destroyed in value, and turned into an irreclaimable bog; and this as the necessary result of the work which the Government has undertaken.”
The court, quoting from the Pumpelly case with the authorities therein cited and concluding that the facts found showed a taking, said:
“ It is clear from these authorities that where the Government by the construction of a dam or other public works so floods lands belonging to an individual as to substantially destroy their value, there is a taking within the scope of the fifth amendment. While the Government does not directly proceed to appropriate the title, yet it takes away the use and value; when that is done, it is of little consequence in whom the fee may be vested. Of course, it results from this that the proceeding must be regarded as an actual appropriation of the land, including the possession, the right of possession, and the fee; and when the amount awarded as compensation is paid, the title, the fee, with whatever rights may attach thereto — in this case those at least which belong to a riparian proprietor — pass to the Government, and it becomes henceforth the full owner.”
With some differences of opinion as to the basis of the conclusion reached on another question, the majority of the court concurred in the reasoning and conclusion on the question of a taking, and it is, of course, the majority opinion which determines the law of the case for us. It is, however, permissible and possibly not without profit to notice the basis of a dissent by the then Chief Justice and two Associate Jus*136tices, as stated in an opinion by Mr. Justice, now Chief Justice, White.
Upon a question of fact, issue is taken upon the conclusion of the court that the findings show that the obstructions placed in the river below the plantation caused the water to overflow the embankment, and it is concluded that the findings show a seepage and percolation of the water through the embankment by reason of the maintenance of an increased height of water against the embankment, causing a superin-duced addition of 18 inches of water on the plantation, and that the necessary conclusion from the findings is that the permanent damage to the property was due to the fact that the drainage thereof into the river had been rendered impossible by the raising of the water level in the river above the outlets of the canals.
Concluding that if damage, by the loss of drainage into the river at mean low tide, was caused by the lawful exercise by the United States of its power to improvement navigation, it was danrmu/m absque injuria and redress must be had at the hands of Congress and not by means of a judicial ruling that a damage so resulting constitutes a taking, the learned justice proceeds to consider the question on the hypothesis that the Government work caused the land to be overflowed by raising the water above the embankment, and says:
“ I do not conceive that there would be a taking, even in that case, of the property, for a remedy would be easily afforded for any permanent injury to the land by raising the embankment. The quantum of damages would thus not be the value of the property, but the mere cost of increasing the height of the embankment so as to prevent the water from flowing over it.”
Arguing that a contention that the payment by the United States of a sum of money necessary to raise the level of the embankment so as to prevent overflows would not compensate the owner because the property would still be worthless for want of drainage, is but to admit that the damage complained of results from the inability to drain the land, which does not constitute a taking, the dissenting opinion concludes:
“ For injury to the drainage the remedy would be readily afforded by, if possible, draining the plantation elsewhere *137than into the river, or by resort to the pumping appliances necessary to lift out the water accumulating from rainfall or percolation. The cost of doing these things would then be the measure of damages. That a resort to these simple expedients is unavailing as to this particular property because of its being situated below high-water mark does not, I submit, show that the Government has taken the property for public use, but simply establishes that the property is so situated that it is subjected to a loss necessarily arising from the fact that it is below high-water mark and therefore absolutely dependent for its drainage on the right of the owner to exact that the mean low tide of the river should be forever unchanged. As the right to so exact does not exist the loss of drainage does not constitute an appropriation of the property by the United States and is but the result of the natural situation of the land. If equities exist, Congress is alone capable of providing for them.”
It may be observed that in the majority opinion, referring to the Pumpelly case, it is said—
“ And on the argument it was conceded by the learned counsel for the Government (and properly conceded in view of the findings) that so far as respects the mere matter of overflow and injury there was no substantial distinction between the two cases.”
Is there not room then for the assumption that but for the controverted construction put on the findings with reference to the overflowing of the embankment and with that feature of the case eliminated the majority holding on the question of a taking might have been otherwise?
In the case of Bedford v. United States, 192 U. S., 217, the facts of which it is not deemed necessary to state, the cases of Pumpelly and Lynah, relied on by the plaintiffs, were distinguished. They are referred to as cases in which “ there was an actual invasion and appropriation of land as distinguished from consequential damage,” and concluding that: “Therefore, the damage to appellants’ land, if it can be assigned to the (Government’s) works at all, was but an incidental consequence of them.” Recovery was denied.
In United States v. Grizzard, 219 U. S., 180, a strip of plaintiff’s land on Tates Creek, tributary to the Kentucky River, was flooded by reason of the erection of a dam in the river which backed up the waters of the river and also those *138of the creek. The court below found that 7-¡- acres of plaintiff’s land was submerged and taken, of the value of $750; that in addition an easement of access by way of a county road to Tates Creek pike was taken, and that plaintiff’s land, worth before the act complained of $3,000, was only worth thereafter $1,500, and awarded judgment for $1,500.
In affirming the judgment, referring to the constitutional provision, it is said:
“The ‘just compensation’ thus guaranteed obviously requires that the recompense to the owner for the loss caused to him by the taking of a part of a parcel, or single tract, of land shall be measured by the loss resulting to him from the appropriation. If, as the court below found, the flooding and taking of a part of the plaintiff’s farm has depreciated the usefulness and value of the remainder, the owner is not justly compensated by paying for only that actually appropriated and leaving him uncompensated for the depreciation over benefits to that which remains. In recognition of this principle of justice it is required that regard be had to the effect of the appropriation of a part of a single parcel upon the remaining interest of the owner by taking into account both the benefits which accrue and the depreciation which results to the remainder in its use and value.”
In support of this declaration as to the rule of “ just compensation” follows a quotation from Bauman v. Ross, 167 U. S., 548.
We think it may not be improper to suggest that in Bau-man v. Ross the court was dealing with a proceeding in condemnation where the act of Congress had assumed to require the consideration of certain elements of benefits as well as damages, and that the principle so enunciated in that case is rested in part at least on the authority of another case which was also in condemnation. The requirement that accruing benefits shall be taken into consideration seems sufficient to raise at least a doubt as to whether it was intended that the rule, should apply in this court to the class of cases under consideration involving a taking for which compensation is sought under the Constitution. Its difficulties are manifest because of the difficulty of applying the element of the rule as to benefits, a practice never heretofore followed and an element not considered in any case appealed from *139this court. The rule must, of course, be applied with reference to benefits to the remaining portion of a tract taken as well as damages thereto if it is to retain its equitable balance.
But we are concerned with another feature of the rule to be deduced from the Grizzard case and authorities cited on that point which, if our view with reference thereto be correct, is of more importance in the determination of these cases than that suggested. An element of the rule laid down is that where a part of a tract is taken the award may properly include, among other things, the damages to the remainder of the tract by reason of the use to which the part taken is to be put. Applied to a taking of property for which just compensation must be made the rule must seemingly rest on the theory that a part of the value of the remainder of the tract, measured by its depreciation, becomes a part of the property taken. Application of the principle to a separate tract of the same owner, even though contiguous and equally damaged by the proposed use of the land taken, is denied. The holding is illustrated by reference to the situation of an adjoining owner. Using our own illus-. tration, an owner of a hundred acres of land in one tract of which fifty is taken may have included in and as a part of his compensation for the taking the depreciation in the value of the remaining fifty because of the use to which the fifty taken is to be put, but an owner of another fifty lying in the same relative position and contiguity to the land taken and equally damaged in depreciation of value by reason of the contemplated use of the land taken, has no constitutional fight of recompense. As to the contiguous land the correctness of the rule, aside from its authoritative source, could scarcely be questioned. The reason therefor is found in the holding that as to it, no part of the tract having been taken, it is a consequential damage for which, in the absence of congressional action to that effect, the Government is not liable. Since in the case put by way of illustration the court was discussing a condemnation proceeding, it is of course apparent that there could in such a case be no award to one whose lands were not the subject of the condemnation, but the question is deeper than that. It involves the question of “ just compensation ” under the Constitution, and as we must *140apply it in this court it relates not to condemnation proceedings instituted by tbe Government but to actions instituted by owners to recover compensation for property taken. Adverting to our illustration, tbe owner whose 50 acres bas been taken may come into this court, because of its jurisdiction in actions founded on the Constitution, and sue tbe Government in effect on two counts, first, for the value of tbe 50 acres actually taken and, second, for one-half tbe value of tbe remaining 50, tbe proportion of its depreciation, and therefore the part thereof “ taken ” by reason of tbe use to which the first 50 is to be put, but the contiguous owner can not sue in any form because, as to him, the same damage as that inflicted on the other owner’s remaining 50 is consequential. The distinction, it seems to us, can not find sufficient foundation in the simple fact that one, by reason of the actual taking, has the doors of this court opened to him while the other has not, for although the one may come into this court for his remedy, that remedy is to be such as he is entitled to because of a “ taking ” of his property and it is somewhat difficult to conceive how the same producing cause with the same resultant effect can amount in the one instance to a taking of a part of the property not actually taken and in the other but to consequential damages. Since in the case of the contiguous owner whose property is not taken the damage is concededly and unquestionably consequential, may we not find therein a strong reason why in the other case identity both of cause and effect should be similarly characterized and treated. The cause of the damage to the remainder of the tract, the use to which that taken is to be put, is a cause which is entirely independent of the mere depriving of the owner of the possession and enjoyment of the land taken and does not come within the usual rule in condemnations that the value of the land taken is to be determined by its value as a part of the entire tract.
The case of Jackson v. United States, appealed from this court and affirmed by the Supreme Court without division, 230 U. S., 1, is a case as between which and the instant case there is more of analogy than is found in any of the other cases cited. In fact the effects of the Government’s works on the waters of the river was the same as are found in this case *141and which we are asked to conclude were caused by the dams. The United States had united with State authorities or organizations and individuals in the construction of a system of levees on both sides of the Mississippi River, the United States for the purpose of improving navigation, and the State authorities and individuals for protection against overflow. At certain points where the bluff or highlands approached near to the river, leaving a comparatively small amount of land subject to overflow, levees were not built, but those above and below were connected with the adjacent highlands at their ends, leaving these lands in effect outside the levees. Plaintiff’s agricultural lands involved in the suit were so situated. They were theretofore protected to some extent by a small private levee, but were subject to occasional overflow, to a greater or lesser degree, by the flood waters of the river, not, however, rendering them unfit for cultivation.
The effect of the adopted system of improvement was to confine the volume of water, theretofore escaping through crevasses, to the channel of the river, to increase the velocity of the current, to increase flood heights and prolong their duration, and to cause more frequent overflows, practically destroying the utility of a portion of plaintiff’s lands for agricultural purposes.
This court, predicating its holding largely on the case of Bedford v. United States, 192 U. S., 217, held that the damages were consequential and denied recovery. In affirming this decision the Supreme Court, through the Chief Justice, said:
“The third consideration, that is, the preventing of the overflow of water by work done in the tributaries and the consequent increase in the volume of water in the river, can not be tested from the point of view of individual authority, as the power to so do involves necessarily the exercise of governmental power. We therefore come to consider the proposition in that aspect. In doing so, however, it is to be observed that even if all the previous considerations which we have stated concerning the nonliability to result from building levees, measured by the right of the individual to build a levee to prevent the water of a river from overflowing its banks and destroying his property, be put out of view, and the case therefore in all its aspects be tested by the scope of the governmental authority possessed by the United *142States, the absence of merit in all claims is too clear to require anything but statement. We say this because the plenary power of the United States to legislate for the benefit of navigation and to construct such works as are appropriate to that end, without liability, for remote or consequential damages, has been so often decided as to cause the subject not to be open. It was directly ruled as to work done by the Mississippi Biver Commission in Bedford v. United States, 192 U. S., 217, 225, upon the authority of which case, as we have said, the court below placed its ruling.”
The fact is not overlooked that in the Bedford case, which is followed in the Jackson case, it was said that “ there was no interference with natural conditions” and the works were in fact to preserve conditions theretofore existing, although to do so involved interference with a natural action of the river in making a cut-off which the works were intended to prevent. But the quoted words were a part of those used to distinguish the case from the Lynah case, but not all. And it is not to be concluded that the affirmance of the decision of this court in the Jackson case was because the work done in the making of the improvements in that case was to preserve and did not interfere with natural conditions. The paragraph quoted above from the opinion of the Chief Justice applied to the “ third consideration,” which had already been stated as follows:
“ Third. The performance of work by the United States tending to diminish the .outflow of water from the' river through streams which flowed from it, to the end that a more efficient body of water might remain in the stream for the purpose of accomplishing the deepening of the channel and thus more effectively improving the navigable capacity of the river.”
Diminishing the outflow of water from the fiver by closing streams which flowed from it for the purposes stated certainly did interfere with natural conditions, but whether that be true or not is immaterial and not the pertinent part of that case in its application to this one. The Government, in the improvement of navigation, had the same inherent' right to build dams such as were built in this case as it had to build levees such as were built'in the Jackson case. The question is as to its liability under the Constitution for the *143results. In the Jackson case the effect of the works was to increase flood heights, prolong them for longer periods, and cause more frequent overflows than theretofore, exactly what we are asked to find in this case, and such were held to be consequential damages.
In the Cress and Kelly cases, 243 U. S., 316, to which we must give attention, it is said with reference to the Jackson case, coupled with the cases of Bedford, Scranton v. "Wheeler and Gibson, that—
“ In each of these there was no direct invasion of the lands - of the claimants, the damages were altogether consequential, and the right to compensation was denied on that ground.”
In the first of these cases to which the material parts of the discussion, as far as we are concerned, are addressed, the plaintiff, Cress, was the owner of land on Whiteoak Creek, a tributary of the Cumberland River. The Government erected Lock and Dam No. 21 in the Cumberland River, and by reason thereof 6.6 acres of plaintiff’s land was subjected to frequent overflows of waters from the river and damaged to the extent of one-half its value. The facts found by the trial court do not accompany the opinion, the learned justice who wrote the opinion did not find it necessary to state them in great detail, and a casual reading might give rise to misconception, but to us the case with which the court was dealing and the holding seem quite clear.
The dam in question was of course below the confluence of Whiteoak Creek with the Cumberland River. It served what we can but know was the purpose of all such dams when in operation, viz, to back up the waters of the river and create a “ pool ” and, that object attained, the result was necessarily to back the waters up into any confluent stream within the limits of the pool.
In the court’s preliminary statement of the case it is said that the land in question is subject to frequent overflows of water from the river and later it is said that the findings “ render it plain that this is not a case of temporary flooding or of consequential injury, but a permanent condition, resulting from the erection of the lock and dam, by which the land is subject to frequent overflows of water from the river.” And referring to the principle that “ overflowing *144lands by permanent backwater is a direct invasion, amounting to a taking,” which it is said is settled by the Pumpelly and Lynah cases, in the former of which it is said that there was an almost complete destruction and in the latter a complete destruction, it is held that “ it is the character of the invasion, not the amount of damage resulting from it, so long as the damage is substantial, that determines the question whether it is a taking,” followed by the same quotation last quoted herein from the majority opinion in our resume of the Lynah case, in which it is to be remembered the findings were construed as showing an overflowing of the embankment and an actual invasion by backwater. And thus it seems to us to appear conclusively from the facts stated, the discussion of the question involved, and the authorities cited that the court, in the Cress case, was considering and deciding as to the effects of backwater caused by the dam in the performance of its usual and intended functions. There is no suggestion of augmented flood waters.
The quotations from the authorities cited, together with such comments as have been made in passing, seem to leave but little, if any, room for discussion as to the rule to be applied in the instant case. If it were for us to determine the question without authorities for our guidance and thus establish a rifle, we must inevitably conclude to the extent that the roads in question were injured or destroyed in the manner stated the damages were consequential, and it seems to us quite clear that such conclusion is required as well under the decisions of the Supreme Court. We find nothing in any subsequent decision which serves otherwise than to justify the application of the rule laid down in the Jackson case to this case.
It is true that the works complained of in the Jackson case were of a different character in that they were levees which served to confine the water to the channel of the river, while in this case they were dams built across the channel of the river and interrupting the flow of the water therein, but, under a long line of decisions not necessary to cite, as to the right of the Government to construct such works in or along navigable streams in aid of navigation, it must be unquestioned that the Government, in the exer*145cise of its dominant right, could build one character of structure necessary for the accomplishment of its purposes as well as the other and, that being true, we have only to consider the effect and liability therefor. The practically identical results found in the Jackson case and alleged in this case are striking. They are the results of the exercise of a legitimate governmental function which in that case was held to be consequential damages, and if one authorized governmental act produces consequential damages and not a taking it is difficult to find any basis for applying a different conclusion to identically the same results following another equally authorized act of the same general nature finding its authority in the same dominant power. In addition to the fact that this case seems properly to be controlled by the Jackson case, we find in the application to it of this conclusion no conflict with any other of the authorities cited.
Conclusions reached are of course determinative of the case, but there are other questions which we deem it proper to consider.
Assuming for the purposes of the question that there was in this case a taking, was the action brought in time or are we without jurisdiction by reason of the elapsing of the statutory period? In other words, if, by the construction of these dams the United States took these roads in the constitutional sense, when did the cause of action accrue? ■
There are three views of the question presented: First, that the cause of action accrued and the statute began to run when the dams, or the responsible dam, if but one, were completed and put in operation; second, that the cause of action accrued when the first damage was suffered; and third, that it accrued at some subsequent time when the taking was to be deemed complete. The rule of certainty, in its different degrees, permeates the law, and it would seem that there could be no more reason for its application anywhere than to the end that the courts and litigants may know with certainty when a right of action exists and when, by the statute applicable, it expires.
A cause of action or a right of action necessarily involves as one of its elements some act or omission of the defendant *146in violation of a right of the plaintiff. In this case the cause of action is an alleged taking of property, not the infliction of consequential injuries, by the construction and maintenance of these dams. The idea of a taking seems necessarily to embody in its very nature a specific act which must have occurred at a time definitely to be ascertained. Such is the ordinary significance of the term, and it is inconsistent with the idea that the specific act is to be supplemented by some indefinite and problematic contingency.
Authorities applicable to this class of cases are meager, if not wholly wanting, but we find the rule quite well settled in another class where there are perhaps less weighty reasons for it than here. In G. C. & S. F. Ry. Co. v. Mosely, decided by the Circuit Court of Appeals of the Eighth Circuit (161 Fed., 72), the railroad company to protect its own right of way along the Canadian Kiver built a line of dikes which it was charged deflected the water of the river and threw it forcibly against the opposite shore and inflicted injuries in successive years by washing away plaintiff’s land to a stated amount each year. The dikes were built in 1893. A part of plaintiff’s land was washed away a month after. The action was commenced, in trespass, in 1897, covering damages to that time and an amended petition claimed subsequent damages. Contending that the dikes were permanent structures the statute of limitations was pleaded. The opinion quotes liberally from other cases and from textbook authorities to which we refer without repetition. The syllabus very concisely states the holding as follows:
“ Held that the dikes were permanent structures and that the damages to plaintiff’s land, both present and prospective, were recoverable in a single action, the right to bring which accrued at once when the dikes were completed and the injury commenced, and was barred in three years thereafter.”
One of the cases referred to with approval and quoted from held in slightly more definite language that—
“ Whenever the nuisance is of a permanent character and its construction and continuance are necessarily an injury, the damage is original and may be at once fully compensated. In such case the statute of limitations begins to run upon the construction of the nuisance.”
*147Dealing with the same general class of cases, all actions on the case, the rule is in some instances held to be that the statute begins to run from the infliction of the first injury. The result of the adoption of this rule would be the same in the instant case as if the rule were as first stated, so that it is immaterial, in that respect, which rule is adopted; but on principle we can not concede its application in a case of this kind however much reason there might be for its adoption in the class of cases in which we find it held. To hold in this case that the cause of action accrued and the statute began to run when the first injury was inflicted is necessarily to determine, it seems to us, that the action is for damages and not for a taking.
The plaintiff contends that the cause of action did not accrue and the statute did not begin to run until long after the building of the dams and not until the cause of action was complete and the roads abandoned, and he relies on the case of King v. United States, 59 Fed., 9, a case of alleged taking by overflow, in which it is held that—
“ Although the water in the Savannah River was raised and thrown on the plantation when the dam was built in 1885, the full consequences were not ascertained and realized until 1888, when the plantation was abandoned. In that year (1888) the cause of action was complete. This action, begun in 1893, is within the statutory period of six years.”
This case is entitled to the consideration which we must always give to an exposition of the law by a learned judge, but we venture the opinion that the rule is not well founded and can not be sustained. The finding of fact on which the conclusion is based is as follows:
“ This gradual result was begun to be felt when the cross-tide dam was finished and was experienced in 1885,1886, and 1887, and gradually grew worse, and has existed continuously to the present time. The plantation was abandoned in 1888.”
In G., C. & S. F. Ry. Co. v. Mosley, supra, the action was on the case against a railroad company. The well-fortified holding in that case and in cases cited, applied to the King case, would have protected a railroad company, if the defendant, from a judgment by operation of the statute and *148because of the rule as to the time it began to run; but the United States, exercising a superior and a sovereign right, must yield to a more flexible and a necessarily indefinite rule. It does not respond, it is true, in damages as does the railroad company, but it responds in a different form for the consequences of the same character of act producing, through natural forces, the same results, but performed, however, with the highest authority and solely for the public good. Can it properly be said that when the United States by the erection of a dam takes one’s land and a railroad company by the erection of a dike in fact takes one’s land, although called damages, the results in both cases following an interference with the otherwise natural flow of the water, the beginning of the period of limitation is to be differently fixed because the remedy is in form different? Theoretically, the taking by the Government may result later from the maintenance of the dam and its continued interference with natural conditions, but damages resulting after the period of limitation in the other case must have been anticipated and recovered for in an action brought within the period. To assume a taking by the United States at a period subsequent to erection of its works is to assume a very peculiar and indefinite situation. In the Lynah case, reviewing the authorities, the court said, “ The rule deducible from these cases is that when the Government appropriates property which it does not claim as its own it does so under an implied contract that it will pay the value of the property it so appropriates.” “ Implied contracts are such as reason and justice dictate and which the law presumes from the relations and circumstances of the parties,” and “ an implied contract is one which the law infers from the facts and circumstances of the case,” and “neither an express contract nor one by implication can come into existence unless the parties sustain contract relations, and the difference between the two forms consists in the mode of substantiation and not in the nature of the thing itself. To constitute either one, the parties must occupy toward each other a contract status, and there must be that connection, mutuality of will, and interaction of parties generally expressed, though not very clearly, by the term ‘privity.’ Without this a contract by *149implication is quite impossible.” (Quotations from authorities cited in Words and Phrases.)
The promise to pay in case of a taking is implied, and the obligation imposed is predicated on the prohibition of the Constitution. An intent to take becomes a necessary part of the transaction if contract liability is to result from the taking. This intent may be implied also, but if to be implied it must be on the theory that the Government intends the natural consequences of its own acts. Must not, therefore, the presumption of intention to take be indulged to create an implied contract to pay as of the time when the act resulting immediately or subsequently in the taking is done ?
We find a further objection, which we venture to suggest, to the rule as laid down in the King case. The finding upon which the conclusion is based recites the apparent results of the Government’s works, as manifested each year thereafter, gradually growing worse, and the abandonment of the plantation in 1888, upon which it is concluded that “the full consequences were not ascertained and realized until 1888, when the plantation was abandoned.” In the instant case resultant damages are alleged to have manifested themselves each year, resulting in the final abandonment of the roads as alleged, although in fact only parts are shown to have been abandoned. To apply the rule of the King case to this case is to assume that the plaintiff might have continued to use its roads under difficulties as long as it saw fit, might protect them if it saw fit and continue to use them, or might, at its own pleasure select the time when it should conclude that the damages were such as to justify it in abandoning the roads and determining that the damages “ amounted to a taking ” and that a cause of action therefor had accrued. A cause of action must arise out of some act or omission of the party to be charged and it is a perversion of usual rules to inject into it any element of personal selection on the part of the other party as to when it comes into being as the application of this rule would seem to do.
Upon the ground of uncertainty it is certainly objectionable, and we can find no better illustration of the proposi*150tion than the case in hand. A petition filed May 5, 1911, alleged with reference to the roads in question such serious and continuous interruption of use and enjoyment “as amounts to a taking thereof.” In the original request for findings filed February 24, 1917, we are asked to find that the injui-ies to said turnpike increased each year from 1903 to 1915, and that “finally, in 1915, much of said turnpike became impassable and destroyed as the result of the Government work and had to be abandoned,” and by a later request we are asked to amend this finding to show that “ the taking of this turnpike did not occur until 1911,” and during the taking of testimony in October, 1915, counsel for plaintiff states of record that “ the testimony heretofore taken in this case (1911 and 1912) has shown that this pike was in use and therefore not taken,” and that “ at this hearing (1915) it is shown that it has been taken and abandoned since the last taking of testimony,” subsequent to which statement is the request that we find it was taken in 1911. It may, of course, be said that it is the evidence and not the requests or statements of counsel which will determine when the taking occurred, but if interested counsel can not develop a satisfactory theory, the rules must indeed be difficult of application.
And there is another suggestion to be derived from the facts of this' case. Several roads declared on in the petition in this case were flooded at pool level. As to them, it is conceded the action was brought too late. The same governmental act is declared upon as to all the roads, both those flooded at pool level and those not so flooded. By the application of the rule under discussion we might have not only a multiplicity of actions but different actions under different rules as to limitation, although all predicated upon the same act.
We are of the opinion that the cause of action, if there was one, as to the roads now in question accrued when Dam 14 was completed and put in operation by the filling of the pool, and that action not having been commenced within six years, we are without jurisdiction.
There is yet another matter to be briefly mentioned. While it does not appear that all the roads in question have ever *151been abandoned, it does appear that such abandonment thereof as has occurred, and the abandonment upon which reliance must be placed, was in 1914 and 1915. Counsel for plaintiff, in taking his further testimony in October, 1915, said of record that the roads were in use when former testimony was taken in 1911 and 1912, after the commencement of the action, and have been since abandoned. Perhaps he could not thus confess away a right of his client, but his statement is a recognition of material facts shown by the evidence.
“ Plaintiff, in order to recover, must have a right of action before beginning suit,” and “ a plaintiff can not supply the want of a valid claim at the commencement of the action by the acquisition or accrual of one during the pendency of the action.” Cyc., Vol. I, pp. 740 and 744, and cases cited. Upon this proposition plaintiff could not recover in this action. It was prematurely brought.
Upon all the questions discussed, both of fact and of law, our conclusion must be against the right of the plaintiff to recover, and its petition will be dismissed with a judgment against the plaintiff for the cost of printing in the sum of $352.38.
It is so ordered.
PIay, Judge; BarNEy, Judge; Booth, Judge; and Campbell, Chief Justice, concur.