This is the third appeal of this case to this Court. It is a Federal Tort Claims Act suit brought under 28 U.S.C.A. §§ 1346(b), 2671-2680. The action is brought by the parents of their fifteen months old child, JoAnn Rewis, both in their representative and individual capacities. The gist of the action is the failure of the Medical Officer of the Day at Walker Air Force Base, New Mexico to diagnose and treat the child for aspirin poisoning from which she later died. The undisputed facts are: at a time estimated to be 3:30 p. m., September 4, 1963, while her mother was ill, JoAnn swallowed an unknown quantity of adult aspirin. Without knowledge of this fact, Sergeant Rewis took his wife to the base hospital at about dinner time, leaving the children with a neighbor who gave them their meal. Upon Sergeant Rewis’ return home he was told that JoAnn had eaten her supper, but then immediately after that she started vomiting severely and had a very bad case of diarrhea. He also saw that “she was blue around the eyes and lips” and he noted her “deep breathing.” He then took her to the Air Force hospital, a distance of about ten miles, where after some delay, he saw the same doctor at 10:00 p. m. who had treated Mrs. Rewis. The doctor noted the symptoms except for the “blueness” and made an entry to the effect that the child was “hyperventilating”; her temperature was normal; he made an inspection of her throat, ears and chest, but took no chemical tests. In light of the fact that the child’s mother had been found previously to have a viral infection, the doctor diagnosed the child’s condition as a viral infection. He thereupon prescribed a “decongestant, acti-fied, one teaspoon four times a day, children’s aspirin, and one baby aspirin every four to six hours a dose, which was one and a quarter grain tablet, liquid diet, a cough syrup, benedryl expectorant cough syrup, one teaspoonful four times a day, and Coca Cola syrup for the nausea, one teaspoon as often as need*1204ed.” The father then took the child home and administered the dosages prescribed during the night, during which the child was extremely restless and crying. He took the child back to the base hospital the next morning where, after some delay, Jo Ann was seen by a doctor who immediately called a pediatrician, who thereafter took a blood sample which showed a blood salicylate level of 87.5 miligrams percent. After this discovery was reported to the pediatrician he set in force heroic measures in an effort to save the child, including a transfer by aircraft to San Antonio, Texas, where it was hoped that the presence of an artificial kidney might save her life. The discovery of the salicylate poisoning was too late to save the child and she died at approximately 7:00 on the evening of September 5, after having been taken from the plane in a moribund condition.
Plaintiffs charge that the failure to diagnose the salicylate poisoning at the time of the first examination at approximately 10:00 p. m. on September 4 was negligent and the direct cause of the death of the child.
The United States defended on the double ground that the conduct of the Medical Officer of the Day on the evening of September 4 fully comported with proper medical practice for the time and place involved and that, moreover, the ingestion of the aspirin was of such a large amount as would have been fatal in any event, so that even had there been negligent conduct by the doctor, there was a lack of legal causation of the death.
On the first two appeals to this Court we dealt with the problems of negligence, first concluding in Rewis v. United States, 369 F.2d 595, that the trial court erred in receiving in evidence answers to hypothetical questions which were based on assumed facts which had not been proved in the record, and in the second, that on the proof developed on the trial in the form of expert testimony the trial court's conclusion that there was no negligence was clearly erroneous, and that in fact the Medical Officer of the Day was negligent in failing to diagnose and treat little Jo Ann Rewis for the illness from which she subsequently died. Rewis v. United States, 5 Cir., 445 F.2d 1303. As pointed out in the second decision, since the trial court had not reached the question of causation at the trial level because of its finding that there was no negligence, the case was remanded to the district court for further proceedings, which all parties understood to mean a hearing and determination of the question of causation.
Because this is a non-jury case, and the trial court is the fact finder', it is of course essential that we ascertain from the statement, findings of fact and conclusions of law that the trial court applied the proper legal standard to the evidence dealing with the issue of causation. While it is not at all clear that the trial court did not have the proper standard before it, one statement contained in the first conclusion of law gives us concern lest too strict a standard may inadvertently have been applied to the proof of the plaintiff in this case.
The court stated, concerning the standard or burden of proof of causation: “This burden was formulated by the Court of Appeals for the Fifth Circuit as a requirement that the plaintiff prove to a reasonable degree of medical certainty that this child’s life could have been saved if Dr. Finley had correctly diagnosed her condition on September 4, 1963. Rewis v. United States, 369 F.2d 595, 599 (5th Cir. 1966). To show that the patient could be saved, the plaintiff need not exclude every possible hypothesis that the child would have died, but need only establish by a fair preponderance of the evidence the reasonable medical probability that it would not have died. Rewis, 369 F.2d 603.”
In this Rewis opinion we correctly stated the standard. “The Court, upon the record as a whole, including the testimony of all the witnesses bearing on the subject, is, as the fact finder, required to determine for itself whether a proper diagnosis and treatment on the *1205evening of September 23 (sic) would likely have prevented the loss of this life. See Sentilles v. Inter-Carribbean Shipping Corp., 361 U.S. 107 [80 S.Ct. 173, 4 L.Ed.2d 142].” (Emphasis added). We then pointed out that on this record in the Rewis case the trial court might have found more than a “likelihood” ; that it might even find causation “to a reasonable degree of medical certainty”. It is unfortunate that this dictum was added, but it did not change the standard for proving causation from “likelihood” of survival to “one of reasonable degree of medical certainty”.
In any event, we emphasize that in an action for damages resulting from negligence in performance of health services there is no requirement “that the plaintiff prove to a reasonable degree of medical certainty that this child’s life could have been saved”, but for the negligent conduct of the physician.
Under the laws of New Mexico which control here, the burden rests upon the plaintiff “to introduce evidence to remove the cause from the realm of speculation and to give it a solid foundation upon facts.” Sanders v. Atchison, Topeka and Santa Fe Railway Co., 65 N.M. 286, 336 P.2d 324 (1959).
The trial court concluded that “the plaintiffs have not carried the burden imposed on them. This Court finds that six hours after ingestion it is more probable that JoAnn would have died than that she could have been saved.” 1
Because we conclude that the government expert witnesses based their opinions on unproved assumptions and on statements of fact in conflict with their own documentary evidence, both with respect to the salicylate level that should be used in estimating the peak level in Jo Ann’s system and because of the fact that the misdiagnosis prevented the initiation of even conservative methods of reducing the salicylate content of the blood, all of which were readily available to the physician, we conclude that the judgment for the United States cannot stand. Moreover, because, as found by the trial court itself, there was competent expert testimony that “if the diagnosis was properly made and certain procedures were followed, I would think that the child could be saved”, we conclude that the judgment must once again be reversed, this time for the entry of a judgment finding causation of death from the misdiagnosis and failure to treat the Rewis child.
In a unique sense the government’s case here depended almost totally on a highly qualified pediatrician who had specialized in the diagnosis and treatment of poisoning, the author of a great many articles, thirteen of which were solely on the subject of salicylate poisoning, Dr. Allen K. Done. Because of the use on the third trial of all of the testimony presented on the two earlier trials, it has not been easy to pinpoint the fatal error in the disposition of this case. However, it is clear that the error arises because of the failure of the trial court to relate the basis of the expert testimony to the hypothesis on which it was based. What must be borne in mind is the fact that the failure of the Medical Officer of the Day on the night of September 4 to diagnose JoAnn Rew-is’ illness as aspirin poisoning, meant that nothing was done, either then, or for more than twelve hours thereafter, to eliminate the salicylate either from the stomach and other organs of the child or from her blood into which it was later absorbed.
Dr. Done’s opinion that JoAnn would have been unlikely to survive was bolstered by his presentation of a curve or graph which, as he explained to the Court, enables one skilled in the art who has ascertained the salicylate level in the *1206blood by the simple taking of a blood sample at a known time following ingestion to compute “a rough guide” to the peak level in a patient following ingestion and absorption into the blood. This chart showed that when the blood sample was taken from JoAnn Rewis between 11 and 12 o’clock on September 5 the salicylate level was 87.5 miligrams percent. By using the equation, which neither appellant nor we fault when used for the purpose for which it was designed, (primarily diagnostic purposes, according to Dr. Done) this would indicate that the peak level of concentration in JoAnn was somewhat over 160 mili-grams percent.2 The trial court concluded that it made no difference whether the peak was 160 miligrams percent or 200 miligrams percent. This may well have been due to the trial court’s accepting Dr. Done’s statement at one place in his testimony that at a level of 160 mili-grams percent “In our experience they all died.” This testimony is belied by other testimony by Dr. Done and completely disproved by his own graph showing numerous survivals at or above the 160 level where conservative treatment was used.
While some of the uses to which the chart were put were criticized by appellants,3 and although we agree that it was potentially seriously prejudicial to the appellants for the time of the taking of the blood sample to be misstated by some two and a half hours,4 we conclude that it is not necessary to rule out the use of the chart completely to arrive at a conclusion that the use of it does not support Dr. Done’s testimony. Without it, there is no proof of what the peak level of salicylate would have been had even conservative treatment been commenced immediately. It must be borne in mind that there is no proof as to what amount of aspirin was eaten at 3 :30, and no attention was paid by any witness to the subsequent ingestion of children’s aspirin upon the doctor's prescription late at night.
One troublesome factor of this case is that a study prepared by Dr. Done based upon what he had determined to be the average length of time it takes for aspirin taken at a single dose (a point on which there is no proof in this record) to be totally absorbed into the blood stream was used here as proof that the presence of 87.5 miligrams percent sali-cylate at a certain time equalled proof of a peak level in excess of 160 miligrams percent; this, even though the graph prepared by Dr. Done made certain as*1207sumptions which he himself suggested were either based on averages or were arbitrarily chosen. For instance, he stated that the graph was made on the assumption that the peak was arrived at six hours after ingestion, although he stated that it sometimes takes as long as eight to ten hours. He also stated “It depends on many factors including whether the stomach is full or empty” and in response to the question “And sometimes you say, the peak level is reached early or sometimes later. It is just impossible to say in any given human being?” He answered “It is impossible to say except within very broad limits.” (Emphasis added). Thus it is that, since no one knew how much aspirin Jo Ann had eaten, and no tests were made when she went to the doctor on the evening of September 4, the only proof as to the “fact” that she had consumed enough aspirin to make her survival unlikely even with proper treatment, promptly taken, is the result of the blood test taken shortly before noon the following day.
Dr. Done plotted the point representing 87.5 miligrams percent on his chart or nomogram at a point representing a time between twenty-two and twenty-three hours after ingestion.5 Then by plotting a logarithmic curve (here a straight line) through this point he concluded that the peak level had been approximately 200 miligrams percent. The nomogram cannot be adequately understood without reproduction. It is therefore attached as Appendix 1.6 It must be borne in mind that this chart was devised by Dr. Done to assist a physician dealing with a child the amount of whose intake of aspirin was unknown, but who was shown at a known time after ingestion to have salicylate poisoning, to approximate the peak level by taking a blood sample. The diagonal descending lines indicate the rate of regression from the peak level that Dr. Done’s studies “approximately” indicated would “likely” occur through normal body processes.
The burden of Dr. Done’s testimony, which was borne out by the other expert witnesses who relied on his studies was to the effect that if the peak level of sa-licylate exceeded 160 miligrams percent a person suffering from this condition would be more likely to expire than to survive.
Now we come to the critical defect in the government’s case. The difficulty here is that Dr. Done made it abundantly clear that his testimony was based upon the assumption that the peak level of salicylate in the child’s blood was 160 miligrams percent or over. But this assumption was never put to the witness as a proven fact as a part of the hypothesis upon which his opinions were based. The only assumption which was proved with respect to the salicylate level was that at some point between nineteen and twenty-three hours after ingestion there was a level of 87.5 miligrams percent. Based on this proven fact, the witness then drew on his own experience and research and study to do some hypothesizing of his own. Government counsel asked Dr. Done “Doctor, in terms of your opinion, what is the significance of the positioning of the Rewis child on your chart ?”7 Dr. Done answered “Well, if one accepts the stated interval since ingestion and assumes that the blood level salicylate was correct, it *1208places her clearly within the range where fatal outcome is really to be expected. It also places her in an area where, the medical probability is that a fatal outcome would ensue.”
In addition to the fact that “the stated interval since ingestion” cannot be accepted for the purpose of his answer, it is also clear that the answer is based upon further assumptions as to the correctness of the blood level salicylate, which means that the graph could be used as proof of the fact that with the 87.5 figure known at a definite time an accurate reading could be made to prove what JoAnn Rewis’ peak level had been at the critical point, here contended by the government to be at or above the 160 miligrams percent mark.
We need only look to Dr. Done’s own explanation of his nomogram and his testimony at the trial to conclude that the Court could not accept his second assumption any better than it could accept the first. Dr. Done testified, as of course he would have to, that he did not know what the peak level was in Jo-Ann’s case. He undertook to substitute this lack of knowledge by a graph which, as Dr. Done testified, was only a disag-nostic aid.8 When questioned by counsel for appellants as to the entries on the chart, including entries indicating what had happened to other patients suffering from aspirin poisoning, Dr. Done conceded that the chart was “minimal in terms of correctness.” This frank answer was to be expected in view of the fact that the assumption that the chart was correct for the purpose of fixing Jo Ann’s peak level at 160 miligrams percent was based on several other assumptions. The first was that the aspirin was ingested in a single dose, whereas we know that at least to the extent that she ingested aspirin after the doctor’s examination at 10:00 p. m., this assumption is not true. Furthermore, there is, of course, no proof as to the actual method or sequence of events when Jo-Ann ate the aspirin. The second assumption which is made in the use of the monogram is that the aspirin had been fully absorbed into the blood stream in six hours. Of course we know this is not true, because JoAnn took additional children’s aspirin after six hours from the first ingestion. Moreover, no such assumption can be made because Dr. Done testified, and his article bore out the statement, that the rate of absorption was affected by the condition of the particular patient, by the kind of aspirin, by the extent to which it may have impacted in the stomach or may have been flaky and easily absorbed and although Dr. Done testified that the time of absorption might be as long as eight to ten hours and that such time could not be stated “except within very broad limits.” The third assumption is that the rate of regression (elimination from the blood) during the hours up to the taking of the blood test the next day was “average”. There was no basis for Dr. Done’s determining, and he made no effort to testify to the effect, that Jo-Ann’s rate of regression was average.
Thus, we do not have an expert witness testifying in response to a hypothetical question in which he is asked his opinion as to the likelihood of survival of a child in whom it has been determined that there was a peak salicy-late level of 160 miligrams percent or higher. He was asked about a child whose salicylate level at a stated time following ingestion was 87.5 miligrams percent, and he thereupon assumed several unproven facts to arrive at the conclusion that she had a peak level of 160 miligrams and that above this figure death was more likely to occur than not.
As we stated in Rewis I “Of course, in order for the opinion to have any value it must be based on assumptions which the trier of the facts can find to have been proved. See United States v. Cooper, 5 Cir., 277 F.2d 857; International Paper Company v. United States, *12095 Cir., 227 F.2d 201.” (Emphasis added). Here there was no opportunity for the trial .court to make a finding that Jo Ann’s peak level was equal to or above 160 miligrams percent. Dr. Done was not even asked whether he had a sufficient basis to form an opinion as to her actual peak level. He did testify that he could not state what it was. Of course Dr. Done did not see the child.9 His testimony was based entirely upon assumed facts and since, as we have said in the cases already cited that an expert opinion is no better than the hypotheses upon which it is based, we must conclude that Dr. Done’s opinion cannot be considered as a basis for the trial court’s judgment.
It is apparent that Dr. Finley, the Medical Officer of the Day, and Dr. Ageloff, the pediatrician who first learned of the 87.5 miligrams percent blood level the following day, based their opinions as to causation equally upon Dr. Done’s logarithmic chart. Each of them discussed the finding of 87.5 mili-grams percent as the basis for his conclusion that no treatment would have been likely to have saved the child. We conclude, therefore, that their opinions as to causation cannot be relied upon to support the trial court’s judgment.
On behalf of the plaintiffs, Dr. Shepherd testified to a number of measures that should immediately be taken upon the ascertainment of aspirin poisoning. He testified as to the importance of the same conservative means described by Dr. Done in his articles. He included in his recitation also the use of exchange transfusion or the use of the artificial kidney for dialysis, mentioning the artificial kidney of course only in a situation where it was available. He then stated “All of these are measures that can be taken and if done early and expeditiously there is every reason to believe that you will succeed in the salvaging of this child, saving this child.” After being queried by government counsel as to the degree of severity represented by a salicylate level of 87.5 miligrams percent nineteen and a half hours after ingestion he agreed that this would be considered “a very severe case.” He was then asked “However, if this had been done, this blood exchange, at six hours?” He answered “That’s an entirely different question. The probability would be far greater for the salvage of life.” (Emphasis added).
Because of the fact that this case has been tried three times, and depositions were also taken, it is not easy to winnow out the hypotheses on which each question was presented to the experts. Nevertheless, we have attempted carefully to analyze the real basis attributed by each expert witness for the opinion given by him. We conclude without any hesitation that Dr. Shepherd’s opinion was based on facts proven in the record. We are equally certain that the opinions of the three government experts were based on assumptions that were not. so proven.
However, because of our reluctance as an appellate court to reverse findings of fact by the trial court, we think it appropriate to discuss a further feature of the ease as it was presented. It is plain that all of the doctors testifying agreed that if there had been a proper diagnosis made at 10:00 on September 4 some immediate measures would have been undertaken. Efforts would have been made to remove the remaining amounts of aspirin that had not yet been absorbed, and of course no additional aspirin would have been prescribed. Other measures would have followed immediately to hasten the removal of the salicy-late from the blood after absorption. From the testimony by the government’s witness, Dr. Done, some of these measures would have caused a rapid decrease in the level.
For some unknown reason, at one point in his testimony as to the patients *1210appearing above the 160 miligrams percent line, Dr. Done said “This is the line above which fatalities are likely to occur In our experience they all died.’’ (Emphasis added). A most casual glance at this nomogram proves this to be demonstrably incorrect. There are several on or above the 160 miligrams line, who received only conservative treatment, and who survived.10 A more significant thing, however, is that clustered just at or below the 160 miligrams percent line there is an even larger number, thus emphasizing the importance of what has been said before to the effect that had the doctor in charge made the examination and given the treatment which proper medical practice would have required, it appears almost as a matter of course that this patient would have fallen among this large grouping under the 160 line, all but one of whom survived.
What we have just said has even greater significance in pointing out why it is essential that the court scrutinize as we have the basic assumptions that Jo Ann’s peak level was 160 miligrams percent. If, because she did not fit into the neat pattern of those (1) whose absorption was completed in precisely six hours, (2) that the ingestion of aspirin was all in one dose, (3) that she was normal with respect to the matter of absorption and (4) average with respect to the degree of regression, there is at least an equal chance that she would have fallen below the line indicated, in which event she would have shown up on the chart as one patient who survived.
We do not reach the question, much discussed by the trial court, as to the relative efficiency of treatment by exchange transfusion and dialysis by use of an artificial kidney back in 1963. It is perfectly clear that Dr. Done gave his opinion with respect to the likelihood of JoAnn’s survival based on the assumption, included in the government’s hypothetical question, that no kidney machine was available. On one occasion he himself suggested that the hypothesis be inserted in the question put to him by counsel.
Then, however, Dr. Done expressed the opinion that, although the process of exchange transfusion was readily available at Walker Air Force Base and although if undertaken it would rapidly reduce the salicylate level, it would have been medical negligence to have undertaken an exchange transfusion, if an artificial kidney was available. The trial court also felt it appropriate to quote from this testimony which, we consider to be totally without value in view of the fact that no artificial kidney was available. The court’s recitation of Dr. Done’s statement “That anyone who had access to hemodialysis would, in his view, be foolish to resort to exchange transfusion because it would simply delay the more effective procedure”, thus turns out to be meaningless, because the Rewis child did not have access to hemodialysis.
The witness’s final opinion, that either with or without a kidney machine he was of the view that JoAnn would have been more likely to have expired than to have survived, makes unnecessary any comparison as to the effectiveness of such treatment. Since, however, this opinion was based upon assumptions not in the record, as we have pointed out above, the opinion is no aid to the government in meeting the prima facie case made by the plaintiffs.
Finally, we think it appropriate to add that Dr. Done’s testimony and his articles, freely drawn from in the discussion of his testimony in open court, clearly and plainly expressed the view that upon diagnosis of salicylate poisoning something should be done immediately to ameliorate the condition. Since it was misdiagnosis of the government doctor which resulted in a failure to take any of these steps and caused the administration of additional aspirin it seems a heavy burden, indeed, to require *1211a plaintiff to prove that if the doctor had done what proper medical practice would require it is more likely than not that the child’s life would have been saved. This, however, as we have noted, is the burden. We conclude that the testimony of Dr. Shepherd, supported as it is in some material respects by Dr. Done, clearly made a prima facie case of causation under the requisite standard of proof. It is equally clear that when the government’s expert opinions on causation are disregarded, because they are unsupported by a proper hypothesis, plaintiffs’ proof stands unrebutted. This, therefore, required a finding by the trial court in favor of the plaintiffs.
The judgment for the defendant is reversed and the case is remanded to the trial court to enter a judgment for the plaintiffs for such amount as determined by the court.
APPENDIX 1
. Tliis ultimate finding is based on a correct formulation of tbe standard of causation, as discussed above. We construe the finding as meaning that tbe court finds that, bad proper diagnosis and treatment been started six hours after ingestion, it is more probable that JoAnn would have died than that slie could have been saved. That is clearly the import of the court’s other findings and conclusions.
. In point of fact the graph presented by Dr. Done assumed that the blood sample had been taken between lKen1i/-1wo and twenly-lliree hours after ingestion. As we read the record, we conclude that the entire case was tried on the assumption that .ToAnn ate the aspirin at approximately 3:30 in the afternoon. It is clear from the government’s witnesses that the blood was taken from JoAnn before she was admitted to the hospital, shown at one time to be II :50 and at another to be 12:00. Moreover, the hypothetical question put to Dr. Done by government counsel stated the child’s blood salicy-late level “was subsequently determined to be 87.5 miligrams percent at about nineteen to twenty hours after ingestion.” By using the figure from twenty-two to twenty-three hours, rather than the correct figure of nineteen to twenty hours the grapli showed that the peak level of salicylate in JoAnn’s blood was 200 miligrams percent. The graph thus seriously misrepresented the most important fact which formed the basis of all the experts’ opinions. The mistake is< understandable, though none the less harmful to the appellants. The author of this chart may have taken the time the laboratory report was given rather than the time when the blood was taken from the child.
. On the chart introduced in evidence the line ICO miligrams percent, was marked “Lethal”. Such an expression of opinion on the chart is, of course, improper. Incidentally, in his 1960 article, later referred to, Dr. Done called this line “Severe”.
. The importance of this will be seen when it is understood that the graph represented studies which undertook to show the peak level because the rate of regression or dispersal of the salicylate from the blood was at a known rate. In other words the salicy-late level would descend at a constant rate from the peak, so that if a figure of 87.5 miligrams percent were shown at twenty-two to twenty-three hours after ingestion it would of course indicate a much higher peak than if it were shown at nineteen and a half hours after ingestion.
. As pointed out in footnote 2, supra, this was incorrect by some two and a half hours.
. Reference to this chart shows how dramat- ■ ic a difference the use of the longer time interval makes in the picture that purports to portray not only .ToAnn’s peak level, but also her position relative to others whose eases were superimposed on the chart. The prej-1 udicial effect of this was heightened bv the printing of the word “Lethal” at the 160 miligrams percent line and the fact that the 200 line (clotted line in the Appendix) and the “x” identifying JoAnn Rewis were drawn •in red ink, whereas the rest of the chart was black. .
. It is to be particularly noted that the Rew-is child was shown on the chart to be at the 200 miligram line, because the 87.5 mili-grams percent projection was drawn at between twenty-two and twenty-three hours following ingestion, a figure which we have shown to be wrong.
. “Q. This graph is only a diagnostic aid, is it not?
A. (Dr. Done) Tes, it is pointed out in the paper.”
. “Q. However, in this particular case that we have just computed, you can’t say, even from your research, that the salicy-i late level, the actual peak level, was even as high as 160 can you?
A. No.”
. It is quite significant that the trial court placed sufficient weight on this statement to quote it in the court’s opinion.