dissenting.
The Special Master has recommended the entry of a decree that would establish a diversion point in the Rocky Mountains and allow Colorado to divert no more than 4,000 acre-feet of water from the Vermejo River at that point; the diverted flow would make an intermountain transfer to supplement the presently inadequate flow of the Purgatoire River in Colorado. Accretions to the Vermejo below the diversion point, as well as the remainder of the original flow, would be available for the four principal users of the Vermejo River. Those four users are all in New Mexico and, of course, are upstream from the point where the Ver-mejo flows into the Canadian River.
A gauge that is located between the second and third of those four users has measured the flow of the Vermejo since 1916. The average annual flow of the river at that point since 1921 is 12,800 acre-feet; if the highest flow years are eliminated, the average is 10,900 acre-feet; if just the 1970’s, which included especially dry years, are considered, the aver*325age is 8,262 acre-feet. No matter which figure is used, the Master’s findings make it perfectly clear that the supply will remain adequate to satisfy the needs of the first three of the four principal appropriators on the river. Colorado v. New Mexico, 459 U. S. 176, 180 (1982) (hereinafter Colorado I). The critical dispute concerns the impact of the proposed diversion on the fourth — the Vermejo Conservancy District.
As the Court noted last Term, the Special Master’s recommendation rested on “two alternative grounds: first, that New Mexico could compensate for some or all the Colorado diversion through reasonable water conservation measures; and second, that the injury, if any, to New Mexico would be outweighed by the benefit to Colorado from the diversion.” Id., at 181. Neither last Term, nor today, has the Court questioned the legal sufficiency of either of those grounds. Last Term, however, we requested the Master to provide us with additional factual findings; today the Court decides that the evidence does not support either of the Master’s conclusions.
I respectfully disagree with the Court’s treatment of two questions of law as well as with its evaluation of the facts.
The Court carefully explains why it has concluded that Colorado’s proof should be judged by a clear-and-convincing-evidence standard. Inasmuch as this is the standard that the Special Master applied, that explanation is somewhat academic. The more troublesome question is what standard the Court should apply when it reviews 28 pages of detailed findings of fact by the judge whom we entrusted to conduct the lengthy trial in this case.
In the exercise of our original jurisdiction it may well be appropriate for us to make a de novo review of the record. The Master’s report is, after all, merely a recommendation and there is no rule of law that requires us to accord it any special deference. I do not think that it would be appropri*326ate in our original jurisdiction cases to accord the same degree of deference that Federal Rule of Civil Procedure 52(a) directs appellate judges to accord to the findings of fact made by district judges in ordinary litigation. Nevertheless, in my view, the cause of justice is more likely to be well served by according considerable deference to the Master’s factual determinations. The record in cases such as this is typically lengthy, technical, and complex. The testimony and accompanying exhibits are much more difficult to assimilate and fully comprehend from the cold record than in the living trial, and of course we do not have the opportunity to make assessments of the demeanor of the witnesses.
The majority repeatedly states that it cannot “find enough evidence” to sustain the Master’s findings. E. g., ante, at 324. Based upon my examination of the trial testimony and exhibits presented to the Special Master, the majority’s search for the evidence must have been cursory indeed. On its face, the majority opinion does not review the evidence in the case; instead it reviews the Special Master’s findings, and in the process of doing so makes general observations regarding the evidence.1
If the Court gave the Special Master’s report the respect that I regard as its due — rather than merely paying lip-service to a “tacit presumption of correctness” ante, at 317 — I believe it would reach the conclusion that his recommendation is fully supported by his detailed findings and that those findings are fully supported by the evidence.
*327As The Chief Justice emphasized in his concurring opinion when the case was here last Term, “these two States come to the Court on equal footing.” 459 U. S., at 191. Colorado is not entitled to any priority simply because the river originates in Colorado, and New Mexico is not entitled to an undiminished flow simply because of its first use. Ibid. We must balance the equities of the competing claims as they existed at the time this controversy began. Neither party should be permitted to improve its legal position by making changes in its use of the river’s waters after our jurisdiction was invoked.
Once these principles are recognized, the “remaining questions are largely matters of fact. The evidence is voluminous, some of it highly technical and some quite conflicting. It has all been considered. The reasonable limits of an opinion do not admit of its extended discussion. We must be content to give our conclusions on the main questions and make such references to and comment on what is evidential as will point to the grounds on which the conclusions on those questions rest. As to minor questions we can only state the ultimate facts as we find them from the evidence.” Wyoming v. Colorado, 259 U. S. 419, 471 (1922).
The first of the two alternative grounds supporting the Master’s recommendation is that “New Mexico could compensate for some or all the Colorado diversion through reasonable water conservation measures,” Colorado I, 459 U. S., at 181.
From the outset of the litigation, Colorado has claimed that New Mexico’s use of the Vermejo’s waters has been wasteful and inefficient. Colorado argues that one “fact” it has stressed throughout the litigation is that “a closed stock and domestic water system could eliminate the waste of over 2,000 acre-feet annually.” Brief for Colorado 41, n. 20, 43-45. This fact — which is essentially undisputed — should be “hard” enough even for the majority, and provides irre*328futable support for the conclusion that there was a significant amount of waste in the District when the lawsuit began.2
The Court sidesteps this point, accepting New Mexico’s argument that the benefits of this system should inure solely to the benefit of New Mexico. But New Mexico simply continues to cling to the position that it should not be required to employ conservation measures to facilitate Colorado’s proposed uses, notwithstanding the fact that we explicitly rejected this position last Term, 459 U. S., at 185-186, and in doing so quoted the following language from our seminal decision in this area:
“The question here is not what one State should do for the other, but how each should exercise her relative rights in the waters of this interstate stream. . . . Both subscribe to the doctrine of appropriation, and by that doctrine rights to water are measured by what is reasonably required and applied. Both States recognize that conservation within practicable limits is essential in order that needless waste may be prevented and the largest feasible use may be secured. This comports with the all-pervading spirit of the doctrine of appropriation and takes appropriate heed of the natural necessities out of which it arose. We think that doctrine lays on each of these States a duty to exercise her right reason*329ably and in a manner calculated to conserve the common supply.” Wyoming v. Colorado, 259 U. S., at 484.
New Mexico argues that the “important factor to consider in regard to the closed domestic and stockwater system is the timing.” Reply Brief for New Mexico 23. It appears that before this controversy arose, water users in the area “began discussing the possibility of building a stockwater distribution system that could save the water necessarily lost” by using the open canals, and a cooperative of water users was formed to investigate “possible solutions.” Ibid, (citing N. M. Ex. No. E-3). Although the users apparently recognized and considered the need to eliminate this waste before this controversy began, Tr. 2765, New Mexico did not take any action to eliminate the waste inherent in the District’s 60-mile network of open canals until after CF&I generated this controversy in 1975 by obtaining a conditional right to divert water from the Vermejo River. We will never know if this waste would have been eliminated but for the existence of this lawsuit; we do know, however, that the water was still being wasted at the time this action was commenced.
With respect to the Vermejo Conservancy District — which of course is the only New Mexico user whose water supply might be impaired by the proposed diversion — the Master found:
“At the heart of New Mexico’s water problem is the Vermejo Conservancy District. Whether lack of administration, lack of diligence, lack of resources or lack of ability is the cause, there is little doubt that the District has failed as a water reclamation project and has serious financial and operational problems of its own. (Tr. 164-169). Several of the conservation problems already discussed are present in the District. Furthermore, there is a problem of loss through evaporation in the District’s seven reservoirs. (Tr. 863, 1296-1299). The District has a 32% efficiency to farm headgates and *330an overall system efficiency of 24.6%. (Tr. 2576). New Mexico claims that the District falls middle range in reclamation project efficiencies. (Tr. 1410-1411). However, the existence of other low efficiency systems is not justification for failure to fully develop water sources here. New Mexico argues that Colorado has merely pointed out areas of inefficient water use without making viable suggestions which would reduce or eliminate the inefficiency. It is the opinion of the Master that New Mexico’s inefficient water use should not be charged to Colorado.” Additional Factual Findings 20.
The majority asserts that the “District was quite arguably in the ‘middle range in reclamation project efficiencies,’” ante, at 318 (quoting Additional Factual Findings 20). See also ante, at 319 (“New Mexico submitted substantial evidence that the District is in the middle [range] . . .”). The Master did not find that the District was within the middle range of efficiencies; he simply observed that New Mexico claimed that was so. The majority cannot bring itself to find in favor of New Mexico on this point, and given the evidence on the issue, that is understandable. One expert witness simply stated: “I know of many systems in which the efficiency is in this neighborhood 30 to 40 percent. ... I know of systems who have lower efficiencies simply because they cannot divert the available supply.” Tr. 1410-1411. When asked if he recalled the testimony of another expert that inefficiencies in that range could not be tolerated in the arid area, the witness responded: “I think he mentioned it would be prudent to make better use of the water supply.” Id., at 1411. Other evidence was offered by New Mexico in support of its claim that its efficiency was in the middle range, id., at 2720-2722, but the methodology of this evidence was highly questionable, id., at 2730-2746, and one expert testified that the District was “extremely inefficient” and “less efficient than any system in Colorado with which I’m familiar.” Id., at 2576. It was this latter testimony that the Master credited *331in explicitly holding that the overall efficiency of the District was 24.6%, implicitly rejecting New Mexico’s position. In light of all of the testimony, the Special Master concluded that “the existence of other low efficiency systems is not justification for failure to fully develop water sources here.” Additional Factual Findings 20.
Moreover, the Master’s findings plainly identify additional conservation measures that are available to New Mexico. They involve a more efficient management of the entire Vermejo River and all specific improvements at the Conservancy District.
The Master noted a marked contrast between the quality of water regulation and control in Colorado, which routinely monitors and takes affirmative measures to eliminate waste, e. g., Tr. 515-524,3 and that provided by New Mexico with respect to the Vermejo River.
In New Mexico, a Water Master is appointed to administer a district if a majority of the users on the system petition the State Engineer, or the State Engineer may do so on his own. Id., at 2424. A Water Master monitors actual use, assures that uses are beneficial, and takes action if there is waste. There is no Water Master for the Vermejo. Incredibly, New Mexico’s answer to the lack of monitoring is simply the assertion that if one farmer “saw another wasting water the matter would be quickly resolved by the water users. Tr. 2416-2417.” Reply Brief for New Mexico 22. See also Tr. 1063-1064. The New Mexico State Engineer testified:
“Even on the streams that have been adjudicated, we find it is generally not necessary to appoint a Water Master to measure the diversions and to enforce priorities and the water users themselves have generally been able to work these problems out among themselves, thus *332avoiding the onerous Water Master tax they would have to pay and the installation of meters that they would have to pay if they demanded strict priority administration.
“Now on the Vermejo we occasionally have had complaints, ‘Somebody is taking water out of priority, filling the lakes when I’m senior,’ things of that nature, and we have sent people over there, talked to the water users in much the same way as they talk to each other. And I think have been of some assistance to them in resolving the problem among themselves.” Id., at 2416-2417.
The same engineer later insisted: “[W]e do not ignore waste. We don’t ignore unadjudicated uses, that is, unauthorized uses for irrigation or any other purpose,” id., at 2418, but later admitted he simply did “not have the staff to go out and monitor for nonuse.” Id., at 2426. Indeed, with his limited staff, he would not even conduct random spot checks, and instead took the position that if he could not monitor all users for nonuse, he would not check for nonuse at all, though he did leave open the possibility in case of undefined “critical circumstances” which he had “not yet encountered.” Ibid.
New Mexico had never installed any gauges at the state line, and did not assist in the maintenance of the gauges installed by Colorado. Id., at 2432-2433. The New Mexico State Engineer did not know the approximate volume of water entering New Mexico, id., at 2433, was “not prepared to so agree” with projections on the effect of the diversion on the New Mexico users, ibid., and was “not able to agree or disagree” with figures regarding depletions, id., at 2433-2434. He explained that such figures were not necessary for New Mexico’s “administration” of the water rights under the New Mexico Vermejo Decree, because his department administered the decree “[o]nly in the sense of occasional fieldtrips to determine primarily whether any un*333authorized acreage is being irrigated. . . . But we do not administer the priorities and diversion rates adjudicated by the decree.” Id., at 2434. “Who does do that?” counsel asked. The State Engineer responded:
“We talked about that some. There is a working among themselves, a cooperation over there. The people work the problems out among themselves. Occasionally complaining to us. . . .
“So long as they are able to resolve them and live with it, then day-to-day administration of priorities and the rates of diversion is not necessary and not in the public interest.
“It’s costly and it costs those water users when we have to undertake that kind of administration. And I think that gives them some incentive to be reasonably cooperative in working out their problems locally.” Id., at 2434-2435.
The problems with relying on complaints by other users are numerous and manifest. Of course, other New Mexico users would have little incentive to complain about waste by the most junior appropriator who in this case is farthest downstream — any water that reaches the District will simply flow into the Canadian River if it is not used by the District. Moreover, one wasteful user will think twice before pointing an accusatory finger at another user wasting water. Naturally without meters and without access to the other users’ land, few complaints are likely. The New Mexico Engineer conceded some of these problems, but simply asserted that the District users “have a pretty good idea what is going on upstream particularly.” Id., at 2424.
In his additional factual findings, the Master specifically suggested the manifest deficiencies in New Mexico’s administration could be remedied by “monitoring, regulating and controlling the system in an effort to determine more accurately actual use, and to decrease nonuse, waste and general *334inefficiency.” Additional Factual Findings 18.4 There is clear and convincing evidence to support the Special Master’s findings and Colorado’s argument that “by means of lax administrative practices, New Mexico precludes a determination of precise demand and actual beneficial use.” Brief for Colorado 41.
Colorado is correct when it states: “New Mexico should not be permitted to use its own lack of administration and record keeping to establish its claim that no water can be conserved. That position, if accepted by the Court, would encourage states to obscure their water use practices and needs in order to avoid their duty to help conserve the common supply.” Id., at 42. Last Term we explicitly rejected New Mexico’s inflexible interpretation of the doctrine of equitable apportionment under which priority would not merely be a guiding principle but the controlling one. 459 U. S., at 183-184. We further stated:
“Our prior cases clearly establish that equitable apportionment will protect only those rights to water that are ‘reasonably required and applied.’ Wyoming v. Colo*335rado, 259 U. S. 419, 484 (1922). Especially in those Western States where water is scarce, ‘[t]here must be no waste ... of the “treasure” of a river. . . . Only diligence and good faith will keep the privilege alive.’ Washington v. Oregon, 297 U. S. 517, 527 (1936). Thus, wasteful or inefficient uses will not be protected. See ibid.; Nebraska v. Wyoming, [325 U. S.], at 618. Similarly, concededly senior water rights will be deemed forfeited or substantially diminished where the rights have not been exercised or asserted with reasonable diligence. Washington v. Oregon, supra, at 527-528; Colorado v. Kansas, 320 U. S. 383, 394 (1943).” Id., at 184.
New Mexico’s manifestly lax, indeed virtually nonexistent, administration of the Vermejo surely substantially diminishes its rights to the waters. It invites waste, and renders the amount of that waste an unknown. “Protection of existing economies does not require that users be permitted to continue in unreasonably wasteful or inefficient practices.” Id., at 195 (O’Connor, J., concurring).
Moreover, the Special Master identified further specific problems causing water shortages or loss that might be alleviated by more careful administration:
“One such problem is unregulated stockponds, fishponds and water detention structures. (Colo. Ex. Nos. 83, 40). While there is no question that such water use is to a certain extent necessary and beneficial, some sort of restrictions should apply. The numbers of ponds and other structures might be limited; when appropriate, reuse should be developed; and, the extent of water diverted to these areas should be in some way monitored or controlled. There is some indication by New Mexico that approximately 2,024 stockponds exist in Colfax County. (Defendants’ Brief on Remand, p. 53). Reduction and/or regulation of some type could not help but *336be an effort, however small, to conserve the water supply and put it to beneficial use.
“There is at least some evidence in reports from the Bureau of Reclamation that available runoff is not being diverted because dams and supply canals are blocked with silt and other debris. (Colo. Ex. Nos. 38, 40, 43; Tr. 2200). Proper administration would make users aware of the diversion problem and perhaps the state and its users together could find means to clean up the canals and prevent further clogging.
“Another problem contributing to water waste and inefficiency is the inability to control headgate spills, divert all the water available, and fully develop all available stream sources. (Tr. 1830-1834,1913-1914). Perhaps repair or revision of the necessary structures is all that is needed, or perhaps resort to a project of more complicated construction is necessary. The Master does not mean to suggest that burdensome and unreasonable efforts are required to be undertaken by New Mexico; however, reasonable repair based on careful development and administration could further reduce water shortages caused by inefficiency and waste.” Additional Factual Findings 18-19.
Based on his review of the entire record, the Master found:
“The Master is of the opinion that based on the evidence in its entirety, there is already sufficient water if New Mexico would take every opportunity to develop their resources fully. With proper conservation measures, there is an adequate water supply to satisfy the needs of all users.” Id., at 20-21.5
*337H-l t — I l — I
Alternatively, the Master found that the benefit to Colorado from the diversion would outweigh the injury, if any, to New Mexico. The identifiable benefits to Colorado included projected permanent uses, interim uses, and the alleviation of the existing shortages in the Purgatoire River system.
The Master found that the proposed permanent uses include
“a water powered hydroelectric plant generating power for a sawmill and related timber operations; coal washing at CF&I coal mines which would save transportation of the waste material from the mines to Pueblo, Colorado as well as development of additional coal mines; domestic and recreational purposes; possible synthetic fuel development; and, supplementation of current inadequate water supply in Colorado, including both CF&I uses as well as city and conservancy district (irrigation) shortages. (Tr. 738-749, 795-96, 623-639, 654, 656).” Id., at 22.
The Master properly acknowledged that there could be no certainty that all of Colorado’s proposed uses would actually materialize, but he concluded that “if even half of them are fully implemented,” the diversion would be justified. He added:
“One of the more important uses, which is certain to occur, is that the water appropriated from the Vermejo *338River will supplement the existing insufficient water supply available to Colorado users. There seems to be little doubt that the Purgatoire River system is over-appropriated, demand exceeding available supply. Any additional water would help to relieve shortages. CF&I and the city of Trinidad are but two examples of users that would benefit by having water available to meet their demands. (Tr. 535-538, 623-630, 795-796). There is some thought that the benefit of alleviating these shortages is sufficient to justify Colorado diversion of Vermejo water; however, Colorado’s proposal does not stop with alleviating shortages but goes on with major plans for the water and thereby additional benefits.” Id., at 23-24.
With respect to the interim period pending full development of permanent uses, the Master found:
“Colorado proposes to temporarily use the diverted Vermejo River water for irrigation of 2,000 acres of agricultural land owned by CF&I. Plans to use and reuse the water as it flows down the valley result in a high efficiency expectation. (Tr. 744-746).” Id., at 22.
The Master credited evidence adduced by Colorado estimating that for its proposed agricultural uses of the diverted water “the efficiency will be 60-75%.”6
The Master again emphasized that reasonable conservation measures “would reduce New Mexico’s doss’ to insignificance.” Id., at 27. He also noted that the District received a significant supply of water from the Chico River, that it has *339four large reservoirs that give it “great ability to store water and enhance the supply,” id., at 127 and, as the Court recognizes, ante, at 318, the District has historically used less water than was available to it.8 Finally, the Master summarized his conclusions concerning the District by stating that “shortages resulting from [the] Colorado diversion (if they exist at all) would be experienced in a project that has failed from the beginning to develop its allotted acreage, has failed to meet its financial obligations, and quite possibly should never have been built.” Additional Factual Findings 8.
> HH
The Special Master’s task was not to draw up blueprints for New Mexico to eliminate its waste. The Master, based on all the evidence, concluded that reasonable conservation efforts in New Mexico would offset the effects of the Colorado diversion. Cf. Wyoming v. Colorado, 259 U. S., at 486 (“Our belief gathered from all the evidence is that, with the attention which rightly should be bestowed on a problem of such moment, it can be successfully solved within the limits of what is financially and physically practicable”). My examination of the testimony persuades me that that conclusion is supported by the record.
Accordingly, I respectfully dissent.
The majority does make a vague reference to certain economic studies commissioned by New Mexico. Ante, at'322. It is unclear, however, whether the majority actually relies on the substance of this evidence at all. Instead, we are told that New Mexico has “attempted to identify harms that would result” and has taken “concrete steps toward addressing the query this Court posed last Term.” Ibid. It seems to matter little whether New Mexico has failed in this regard, because its analysis is “no more speculative” than Colorado’s evidence. Ibid. The majority nevertheless gives New Mexico an “A for effort,” as it were, whereas Colorado is seemingly penalized because it “has made no similar effort,” ibid.
Colorado further argues that the diversion it seeks would be totally offset by this savings. The argument is based on the fact that the saving of 2,000 acre-feet is realized at the reservoirs in the District, and that there is a significant loss of water during its transit from the river to the reservoirs, and also resulting from evaporation from the reservoirs. Thus, according to Colorado, an increase of 2,000 acre-feet of water in the reservoirs would offset a much larger diversion from the river itself. One need not fully accept this argument to recognize that the recommended 4,000 acre-feet diversion upstream would produce a significantly lower net loss at the reservoirs, or — more significantly — that when the complaint was filed, at least 2,000 acre-feet of water were being wasted by just one of the four principal users in New Mexico.
It was in light of this evidence that the Special Master stated that “it is not for the Master or for New Mexico to say that reasonable attempts to conserve water will not be implemented by Colorado.” Additional Factual Findings 21. See also id., at 14-16.
The Master further stated:
“One final problem area which the Master believes could be improved with proper administration is the failure of many users to devote sufficient time to the complete development of available water resources. Water shortages are a reality in arid western states and, therefore, water conservation is a task that must involve serious effort and attention together with large amounts of time and financial input. The Master understands the intense feelings that some of the individual users have for their land and their lifestyle (See Tr. 2192, 2206, 2215-16); the Master also understands that farming or ranching often needs to be supplemented by other sources of income and, therefore, other jobs. (See Tr. 2207). However, New Mexico users, individuals, or otherwise, cannot expect to be able to take the available water in the Vermejo River at their convenience without taking the time and energy to implement changes and development to help conserve and augment the available water. Careful monitoring and regulation as part of a program of administration would aid all users in full development of their water supply and demands.” Id., at 19-20.
In the conclusion of the report the Master expressly stated:
“The available supply of water from the Vermejo River is sufficient for current New Mexico users, and with reasonable conservation measures would meet the needs of Colorado users as well. The available water supply can be enhanced through diligent and complete development of the Vermejo source as well as alternative sources. Many current users do not *337require a continuous supply and systems of reservoirs provide relief for those who do.” Additional Factual Findings 28.
While Colorado did not undertake a detailed study of ways to improve the efficiency of the Vermejo system in New Mexico, thinking that it was not its place to administer the Vermejo in New Mexico, Tr. 238-239, based on the evidence available, its experts concluded that reasonable conservation measures would offset the diversion, e. g., id., at 243, 247, 876, 2579. This expert opinion testimony was plainly admissible on this ultimate question, Fed. Rules Evid. 702, 704, and together with other evidence in the record, fully supports the Master’s conclusion on this question.
“There is no reason to doubt the validity of Colorado’s proposals or intentions. Even if the actual does not comport with the ideal, it is not for the Master or for New Mexico to say that reasonable attempts to conserve water will not be implemented by Colorado. The strict administration of water already on display in Colorado increases the likelihood that the proposed measures will be implemented at least to a reasonable degree.” Additional Factual Findings 21.
See also id., at 27 (“As noted earlier, the District has a reservoir system allowing carryover from wet years to supply water during periods of shortage. Therefore, the user most affected does have a means of offsetting the possible shortage”).
The District has irrigated an average of 4,379 acres although it has rights from the Bureau of Reclamation to irrigate 7,979 acres. Id., at 8. Moreover, the Master found that two individual farmers with water rights senior to the District, but whose farms are located downstream from the District, have historically used less than their decreed rights even though the supply was adequate to enable them to develop their entire acreage. See id., at 6-7.