MURRAH, Circuit Judge.
These consolidated cases involve claims for damages to livestock in Salt Lake and Utah Counties, Utah, allegedly caused by the emission of gaseous and miniscular [424]*424particles of fluorine from the United States Steel furnaces at Geneva, Utah. The three hundred claims of livestock owners involving common questions of law and similar facts, were grouped into these cases and others unappealed.
Each of the claims charged that since 1950, the appellee had knowingly permitted large quantities of fluorine gases to be emitted from its Geneva plant; that these gases, borne by prevailing winds, contaminated the vegetation growing upon the respective appellants’ land; and that such vegetation was consumed by the livestock in sufficient quantities to cause a disease called fluorosis, with subsequent physical damage to all of the livestock and milk loss from the dairy cattle. Injunctive relief against the operation of the plant was originally sought, but during the pendency, of the litigation, corrective measures reduced the rate of fluorine emission to the point of admittedly harmless result. Injunc-tive relief was thereupon abandoned, but each of the separate claims for damages state the jurisdictional amount in controversy and requisite diversity of citizenship.
Issues having been joined on liability and extent of damage, if any, the trial court sitting without a jury heard 260 witnesses, including the claimants and experts for both the plaintiff-appellants and appellee-defendant. Reports of some 100,000 chemical analyses of vegetation for fluorine, and diagnoses of about 12,-000 cattle and thousands of sheep were received in evidence.
At the conclusion of all the evidence, the trial court by agreement of the parties appointed a medical doctor and former Dean of the University of Utah School of Medicine, as a neutral expert and referee, to “marshal the evidence [from 6,622 pages of testimony and 1,-267 exhibits], discuss it with counsel, case by case and cow by cow, hear them out and report his findings to the court.” And, thereafter, as the court stated, “counsel presented ‘brochures’ containing 10,000 pages summarizing the evidence for the referee, setting forth their respective views supported by extensive briefs and months of oral argument.” The referee’s report in the several cases, consisting of 347 pages, was adopted and incorporated by the trial court in its extensive opinion, in which it directed the entry of judgment generally for the defendant, but granted recovery to some of the claimants in varying amounts. At the same time, the court entered findings of fact in substantial conformity to the referee’s recommendations, which findings, together with conclusions of law were in accordance with those previously suggested by the appellee.
From the scientific proof, the referee found, and the appellee concedes, that during the years complained of, potentially harmful quantities of fluorine gases did emanate from the appellee’s Geneva plant; that it did fall upon the appellants’ lands and vegetation in varying quantities; and that such vegetation was consumed by the appellants’ livestock. And, the appellee further concedes its legal liability for any substantial harm caused thereby. And see Reynolds Metals Co. v. Yturbide, 9 Cir., 258 F.2d 321; E. Rauh & Sons Fertilizer Co. v. Shreffler, 6 Cir., 139 F.2d 38; Anderson v. American Smelting & Refining Co., D.C., 265 F. 928.
Based upon controlled experiments and other scientific analysis, the referee established a “tolerance level” below which the ingestion of fluorine by livestock was found to be harmless and above which there was a possibility, and then a probability of harm. Specifically, the referee and the trial court found that, with respect to cattle, continued ingestion at the rate of 30 parts fluorine per million parts total intake was below the tolerance level; that 30 to 50 ppm created a possibility of harm; and, levels above 50 ppm, injury was probable. The critical level of continued fluorine intake for sheep was found to be 70 ppm. From evidence of scientific forage sampling and atmospheric tests, the referee suggested, and the trial court found, that in a number of areas involved, fluorine in forage reached the border line level of [425]*425harmful concentrations. And, it was therefore necessary to consider the available data relating to specific levels of exposure in connection with the claims of each claimant.
In addition to the forage exposure levels, the scientific proof developed reliable diagnostic aids in determining the damaging effects of fluorine intake on livestock. The first, and most sensitive, is condition of teeth erupting during exposure. The teeth were generally classified as numbers 0 to 5, and it seems to be generally agreed that the teeth classified as 1, 2 and 3 do not indicate any harmful effects, while teeth classified 4 and 5 indicate some fluorine injury. Second, the fluorine content of the hones, which may gradually build up during periods of excessive fluorine ingestion, with symptoms of fluorosis appearing at levels of 4000 to 5000 parts fluorine per million content. Third, urinalysis with 5 ppm as the normal level and 10 ppm as an elevated fluorine systemic content, and consistent levels above 25 ppm suggesting definite fluorosis.1 Fourth, clinical symptoms, such as lameness, poor nutrition, small stature, rough hair coat, tight skin, loss of milk production, lapping of water, and reproductive difficulties. These clinical conditions are generally considered secondary or non-specific evidence of systemic fluorosis. And, it is generally conceded such conditions may arise from other and independent causes. Though it seems to be well established that none of the diagnostic aids are in and of themselves conclusive2, the parties apparently accept these combined factors as standards for the determination of fluorine injury to livestock. Indeed, the burden of the appellants’ contention on appeal is that the referee and the trial court utterly failed to apply these accepted standards in determining' damage to the livestock of the respective claimants. We look, therefore, to the referee’s consideration of livestock injury in the respective herds of each group of cases to determine whether his report and the findings of the court thereon are compatible with these standards and substantially supported by the record. In consideration of this decisive issue, it seems almost superfluous for us. to reiterate the time-honored and oft repeated canon for the interpretation of Rule 52, F.R.C.P., 28 U.S.C.A., to the effect that in determining whether the judgment of the trial court is reversibly erroneous, we will not weigh the evidence or judge the credibility of witnesses. The appellate function is to determine whether the challenged judgment finds support in the facts and warrant in the law — in sum, whether there is a rational basis for what the trial court did. See United States v. Oregon State Medical Society, 343 U.S. 326, 72 S.Ct. 690, 96 L.Ed. 978; Skinner v. Parnell, 10 Cir., 257 F.2d 345; Fuller v. C. M. & W. Drilling Co., 10 Cir., 243 F.2d 862; Heldenbrand v. Stevenson, 10 Cir., 249 F.2d 424; Shoemaker v. Leeper, 10 Cir., 186 [426]*426F.2d 372; Brown v. American National Bank, 10 Cir., 197 F.2d 911; North Drive-In Theatre Corp. v. Park-In Theatres, 10 Cir., 248 F.2d 232; Levelle v. Powers, 10 Cir., 248 F.2d 774. And, the canon is to be sure cogently applicable where, as here, the judicial process necessarily involves scientific inquiry, and is based upon the opinion of those schooled and trained in the operative facts of the suit. See Millers’ Nat. Ins. Co., Chicago, Ill. v. Wichita Flour Mills Co., 10 Cir., 257 F.2d 93, and cases cited.
Erekson Case (No. 5811)
This case, the first filed and the first tried, involved 16 claimants, owners of dairy farms in Salt Lake County, Utah, at distances ranging from 16 to 26 miles generally north of the defendant’s plant, from which they are separated by a relatively low transverse range of mountains intersected by a river. The court found that these claimants proved no damage, except to two cows which were found to have systemic fluorosis and for which recovery was allowed in the amount of their respective value unaffected by fluorine.
The referee’s report, accepted by the court in its findings, stated that the forage levels of fluorine on these farms were generally below the tolerance limit, but that in a number of areas fluorine in forage reached the borderline zone of harmful concentrations. And, furthermore, levels of fluorine found in urinary tests of some of the cows, bone levels found in laboratory tests of autopsied animals and the results of examination of the cows’ incisor teeth, all indicated above normal intake of fluorine consistent with possible mild systemic fluorosis, but not necessarily leading to that conclusion. Thus, the referee turned to individual diagnoses of animals to determine whether they suffered from systemic fluorosis.
The claimants testified of clinical symptoms such as loss of milk production, unthriftiness, rough coats, settling .difficulties, small stature and poor general condition of their cows. All of these symptoms are admittedly nonspecific, but are undoubtedly indicative of some deleterious condition, which might well be attributable to systemic fluorosis. Veterinarians testifying for the respective claimants made a blanket diagnosis that all the herds of the claimants were affected by fluorosis, and singled out individual cows as being harmfully affected by fluorine. But, as the referee pointed out, none of them offered a single case of systematic study of an identified animal with itemized findings coupled with medical analysis leading to a diagnosis of systemic fluorosis.
The appellee offered the evidence of a veterinarian and a biochemist who, the referee found, had examined every available cow of all the Erekson claimants. These experts made extensive scientific tests, including autopsies of 16 animals. The essence of their composite testimony is that, except for and notwithstanding dental lesions on incisor teeth from fluorine, they failed to find indications on which a diagnosis of definite systemic fluorosis could be made on any cow, though some were considered in borderline status, and two cows (for which recovery was granted) were reported with exostosis due to fluorine. In addition, the defense experts directly refuted the testimony of the individual claimants on clinical symptoms of the general condition of the cows.
The claimants argue that scientific evidence in the record credited by the referee, and common sense as well, indicate that excessive fluorine in an animal’s system will be reduced by excretion, when the animal is removed from exposure. They insist that the high point of fluorine emission was reached in 1951, and thereafter reduced from year to year to a harmless level; that in the greater part, the first examinations of the claimants’ herds by defense experts took place in 1954; and thus no evidence conflicts with their testimony of injury before that time.
Their hypothesis is not aided by the referee’s opinion that urinary tests did [427]*427not show sufficient fluorine content to indicate fluorine drainage from the skeleton of animals later autopsied and found not to have an injuriously high fluorine level in their bones. But the short, complete answer to this contention is that in testifying at trial, neither the claimants themselves nor their veterinarians distinguished between the condition of animals at different times in this case.3
The chief complaint of all of the litigants, and the burden of their proof, was lowered milk production from fluorine, and they sought to recover the value of this alleged loss. Though there was testimony that individual cows had not done as well as they should, the plaintiffs’ general approach was to set a level of expected normal production on a herd basis and seek damages in the amount the herd production did not measure up to this standard.
Defense evidence showed that in the years 1949 through 1951, all of the litigants’ herds outproduced the state average, with the southern tier of six farms, closer to the steel plant and presumably more exposed, having higher production than the northern tier of ten farms. Then in 1952, both groups experienced a reduction to the state average. Thereafter, the northern tier maintained the state average, while the southern tier crept above it. The referee expressed an opinion that over many years these variations might seem to approach normal periodic fluctuations, but that nevertheless some adverse influence on milk production seemed to have been at work beginning about 1951.
The referee, however, considered milk production more a matter of possible damages than an aid in diagnosis of systemic fluorosis. He expressed considerable doubt as to accepting lowered milk production as a result of mild systemic fluorosis, though he stated it would occur in cases of pronounced fluorosis. This opinion of the referee, together with the defense evidence indicating the lack of even mild systemic fluorosis, would seem enough in itself to support the court’s finding that there was no cognizable milk loss. But in addition, Dr. Schmidt of Stanford testified that an examination of the DHIA (Dairy Herd Improvement Association) milk production records of the claimants over a number of years did not indicate to him a lowering of production in the fluorine era of the early and mid-1950’s, except for one herd in which the result might be attributable to management practices rather than fluorine. And the defendant argued that the milk records generally show a higher production per cow on the claimants’ farms in the fluorine era than before. These additional factors certainly serve to protect the court’s finding from attack as clearly erroneous.
In the last analysis, and at most, there was a conflict of evidence and a choice of expertise on the vital issue of fluorine damages. It is sufficient to say that the referee and trial court thought the evidence for the defense the more credible, as we think they well might have.
Abbott Case (No. 5815)
This group of cases involved about 180 claimants whose dairy and beef farms all were located in the same general area of Salt Lake County as those in the Erekson group. The court found that many of these claimants failed to prove injury from the defendant’s fluorine. Recovery was allowed, however, in varying amounts for animals found to be damaged by fluorine. Conditioned upon their delivery to the defendant, recovery was also granted for certain cows found to have molar wear as a result of fluorine which did not yet interfere with the animals’ functions, but which might progress to such a point in the future.
[428]*428In this case the court made the same finding concerning general levels of fluorine in vegetation tested as in Erekson, and likewise considered the evidence of damage to individual cows and herds. In his comprehensive report, the referee recognized that injury to cows from excessive ingestion of fluorine might come about through either systemic fluorosis, in which the fluorine has a direct toxic effect on the vital processes, or through dental fluorosis, in which the molars are so worn by fluorine decay that the cow cannot properly masticate its food when “chewing its cud” and therefore suffers from malnutrition.4 Either condition results in unthriftiness and lower milk production with commensurate reduction in cow value. In the absence of any tests of molar decay in the Erek-son case, the referee had no occasion to consider this important factor in the determination of fluorine injury there. In this case, however, the referee had available, and considered at length, reports on the molar condition of many animals.
Many of these claimants testified of unthriftiness, stiffness, poor coat, small stature, lapping of water, and general estimates of lowered milk production in their respective herds. Veterinarians for the claimants examined practically all of the claimants’ herds, inspecting the incisor teeth of the cows, but not the molars, and further observing the general condition of the cows and such specific indications of fluorosis as apparent bone effect. As in Erekson, it was their conclusion that every one of the claimants’ herds had been harmfully affected by fluorosis. The lack of variation in their reporting of symptoms of severe fluorosis in nearly all of the hundreds of animals examined, and the sweeping generality of the fluorosis diagnosis, led the referee to comment unfavorably on the weight to be accorded testimony of the claimants’ veterinarians.
Though the defense presented some evidence of laboratory tests conducted on certain of the claimants’ cows (particularly fluorine bone level test), the bulk of its evidence consisted of the testimony of its experts who had examined, sometimes twice, available animals in practically all of the claimants’ herds. Except for isolated instances, in which systemic fluorosis was found and recovery granted, the defense witnesses directly refuted the claimants’ evidence of clinical symptoms of fluorosis. And, the referee and trial court accepted their opinion, based on molar as well as incisor examination, that the principal harmful effects of fluorine in litigants’ animals was due to tooth wear, or dental fluorosis.
In general, the court allowed recovery for the full value of any cow found to have severe or marked molar wear. In addition, the court found that certain animals had slight to moderate molar wear resulting from dental fluorosis which had not yet resulted in functional injury to them, but which might some time in the future. To do equity, the court fixed the value of these cows, and decreed that the plaintiff could collect this amount from the defendant upon delivery of the animal in question to the defendant within thirty days after judgment.
This method of conditional recovery is suspect because it granted no award for animals sold or lost between time of diagnosis and time of judgment. The trial court, however, in its opinion stated that this procedure was stipulated by counsel, thus removing any cause for complaint. The appellants claim that they stipulated to the procedure of conditional recovery based on market value of the cows assigned by the claimants rather than the value found by the court, which accepted in the greater part the appraisals of the defendant’s witnesses. But having accepted the form of conditional recovery, the claimants cannot now [429]*429object to the procedure on the grounds that the court based its findings of value on the evidence it deemed most credible, rather than accepting the values set by the claimants.
The appellants charge the court with reversible inconsistency in granting absolute awards for certain cows with moderate molar wear, and conditional recovery for other cows with the same degree of molar impairment. But, consistent with the formula of basing its diagnosis on all of the accepted diagnostic factors in determining harmful fluorosis, the court’s discrimination rested upon clear evidence of difference of clinical condition, with one possible exception, even as to which we cannot say the court clearly abused its discretion.
Though allowing awards, either absolute or conditional, to this group of claimants for cows with molar wear, the court, following its practice as in Erekson, granted no recovery based solely on incisor tooth damage. The claimants complain of this, saying that since molar wear tends to parallel incisor wear, the court should have indulged in the factual presumption that molar damage followed incisor damage. The referee did report that incisor injury indicated molar wear, but not necessarily so. He accordingly again turned to secondary symptoms to determine in each case whether such symptoms supported his ultimate finding of fluorine damage. The referee and trial court evidently chose to believe the defendant’s expert testimony concerning the clinical condition of these particular cows, and any permissible presumption was thereupon rebutted. It follows that the court’s resolution of the critical factual issue is binding here.
The appellants’ main complaint was again loss of milk production. However, in contrast to the Erekson group, only two or three of these claimants kept extensive records on milk production through DHIA dairy testing association before 1952. For this reason, they were able to give only general estimated percentage of lowered milk production, which they felt came about through the deleterious conditions of their cows to which they testified and which their veterinarians attributed to fluorosis.
The defendant adduced evidence from other farmers in the area and a professor at Utah State Agricultural College that variations of milk produced were within the range of normal fluctuation. The defendant also presented evidence that loss of milk, if a result of systemic or dental fluorosis, will be accompanied by other signs of malnutrition or un-thriftiness, which the defendant’s experts failed to find among the claimants’ cows in forms associated with fluorine, except for a few cows for which recovery was allowed. The court found the claimants suffered no milk loss from fluorine.
Again, as in Erekson, the trial court, considering his observation of the witnesses and the referee’s report, resolved a conflict of the expert opinion in favor of the defendant, and we will not disturb his choice. There are, however, some isolated instances of error which we are constrained to notice.
The referee, relying on an autopsy by defense experts, reported that the evidence revealed thereby justified a diagnosis of fluorine systemic injury to cow Merlin II of the Joseph Sander Anderson herd, but no mention of Merlin II was made in the court’s findings. In view of the standards set by the court, we will av/ard to the owner of this animal the sum of $200, the amount of his claim.
The court awarded Elvoy Dansie damages in the amount of $100 for injury to cow Daisy. The undisputed appraisal by the owner shows the value of the cow to be $325 with salvage value of $125. Thus, the award should have been $200, and we increase this claimant’s recovery $100 to reach this amount.
No defense expert examined the cows of V. W. Bently, which he claimed were damaged by fluorine during the time he owned a farm in southern Salt Lake County from August 1952 to July 1954. This was the southernmost farm in the Abbott group, and was considered by the [430]*430referee “possibly the most exposed to Geneva influence.” The trial court granted no recovery to this claimant, despite his testimony of such symptoms of clinical fluorosis as unthriftiness, shaggy hair and dry skin, coupled with the testimony of the claimants’ veterinarian of incisor injury and symptoms of systemic fluorosis. The trial court generally did not display much confidence in the claimants’ veterinarians, but in the absence of any defense testimony concerning this herd, it cannot be said that the claimant’s uncontroverted evidence failed to make a case. Accordingly, we are of the opinion that an award should be made to this claimant in accordance with his evidence. There are certain other herds which defense experts did not examine and on which recovery was not allowed, but in the greater part they did not appeal, and in any event, the evidence concerning them, though uncontroverted, does not show sufficient proof of damage to justify an award.
One last error seems to be a mistake in transposing figures from the referee’s report to the court’s findings of fact, which apparently and understandably occurred because of the length and complexity of the damage considerations within this group of claims. Among other awards to Martin S. Larsen, the court granted recovery for five groups of animals, consisting of Hereford cows, Jersey cows and a Jersey bull, in accordance with the report of the referee. On three of these groups, the court awarded the amount of the claimant’s claim, but on the other two, it awarded recovery in the amount of the animals’ salvage value after fluorine injury, which sums the referee reported immediately above the amount of claim. To be consistent with the other three awards and the evidence, these awards should also be in the amount of the claim, and the trial court undoubtedly so intended. This adjustment will increase the claimant’s total recovery by $1,230.
Johnson, Rudd and Flanders Cases (Numbers 5813, 5814 & 5816)
These three claimants had farms in Utah County. The trial court found against Johnson and Rudd, but awarded damages to Flanders. All three appeal.
On the Flanders dairy farm, about three and one-half miles from the defendant’s plant, the trial court, accepting the testimony of the defendant’s expert witnesses, found eight animals who had suffered sufficient molar wear from dental fluorosis to destroy their functional usefulness, and two cows that were functioning properly but which had sufficient molar wear to indicate that they might some time in the future be damaged. For the former, the court awarded damages in the amount of the value of the cows if uninjured. For the latter, the court awarded the claimants the market value of the cows conditioned on their being delivered to the defendant within thirty days after judgment.
The court awarded no damages for alleged milk loss. This finding was supported by testimony of defense experts that the claimant’s cows did not exhibit the secondary symptoms of malnutrition which would accompany milk loss from fluorosis. Though the referee stated that the degree of molar injury in certain animals might have been sufficient to depress their condition and lower their milk production5 as in other cases, he was unable under the proof to ascertain that any actual and specific damages had resulted. Again we give credence to the trial court’s judgment based on the reasoning of its neutral expert.
Claimant Johnson owns a dairy farm about one and one-half miles from the defendant’s plant. The court found that the claimant suffered no damage as the result of the defendant’s operations. The [431]*431usual expert witness evidence of the defendant received exceptional bolstering in this case from the testimony of a man who leased the dairy from the claimant in 1954 and found the herd in good condition and producing “very well”. Further, the defendant introduced evidence of higher average milk production per cow in 1953 than in 1948.
Claimant Rudd’s farm, about two and one-half miles south of the plant, raises primarily beef cattle with some dairy calves. He testified that his animals did not gain weight as they should. However, there was apparently no evidence of forage level tests for fluoride or individual examination of his cows, though there was testimony about the herd generally from the claimant’s veterinarian to the effect that they were not doing well. The evidence was apparently not sufficient to convince the trial court that claimant Rudd had met his burden of proof of damages, and indeed the evidence is sufficiently sketchy to indicate that the findings of the trial court were certainly not clearly erroneous. In these three cases, substantial evidence supports the findings of the court.
Butterfield Case (No. 5812)
These claims were made by some twenty-seven sheep owners in Utah and Salt Lake Counties, alleging damage from fluorine ingestion to their sheep. Four of these claimants were awarded small amounts of damages by the court, and in the other cases, judgment was entered in favor of the defendant. The four sheep owners who received damages, and four who did not, appeal.
As we have seen, sheep are more resistant to fluorine than cows, but the forage level of fluorine led to examination of the various herds for possible ill effects. The trial court found that a few sheep were injured to a point requiring awards to their owners, but as to all of the sheep but a negligible per cent, the trial court simply found that the claimants had not proved damage to their function from fluorine. Thousands of these sheep were examined by experts of both the claimants and the defendant, and as was its prerogative, the court decided to accept the defendant’s evidence.
The court found, however, that a few sheep in certain herds that foraged at widely separated areas in different seasons had molar wear from fluorine subject to deterioration to a point of interference with the animal’s functions, but that the claimants had not proved the defendant to be the source of the injurious fluorine. There was some evidence in the record of other industrial sources of fluorine in northwestern Salt Lake County and some ambiguous forage level tests, but the totality of the evidence leaves no doubt that the defendant’s plant, at the very least, substantially contributed to the presence of any fluorine ingested in harmful quantities. Accordingly, in the light of the court’s findings of values of individual sheep, recovery should be allowed Alonzo Freeman, in the amount of $51, and Parley Spratling, in the amount of $34.
This brings us to an affirmance of the judgments in all of the cases except with respect to claims and instances noted. As to them, the cases are remanded with directions to enter judgments in accordance with the views herein expressed.