MEHAFFY, Circuit Judge.

This (diversity products liability action)) instituted pursuant to the provisions of 28 U.S.C.A. § 1332, involves a multiple antigen product marketed by Parke-Davis and Company under its trade name “Quadrigen.” This product was intended for use in simultaneous immunization against four major diseases of children: diphtheria, tetanus, pertussis (whooping cough) and poliomyelitis. Shane Stromsodt, a minor, by amended complaint brought this action by Robert M. Stromsodt, his father and guardian ad litem. Appellee will sometimes hereafter be referred to as plaintiff and sometimes as Shane.

Shane was born in Grand Forks, North Dakota, on May 24, 1959 and was inoculated with Quadrigen by a technician under the direction of Dr. John H. Graham, the family physician, on August 26 and October 1, 1959. Following the second inoculation, plaintiff suffered a convulsion and thereafter suffered other convulsions. There is no dispute that he sustained serious and permanent brain damage, disabling him both mentally and physically to the extent that for his life *1392 time he will require constant attention and perhaps institutional care.

The case was tried to the United States District Court for the District of North Dakota sitting without a jury. The trial court found that Quadrigen proximately caused Shane’s condition; that Parke-Davis breached its implied warranty of fitness and merchantability; and that Parke-Davis was negligent in its testing of Quadrigen and in its warning of possible danger from its use. The trial court rendered judgment in the sum of $500,000.00, and its memorandum and order are published at 257 F.Supp. 991. We affirm.

The issues before this court may be broadly stated as: (1) causation (cause of brain damage sustained by plaintiff); (2) sufficiency of the evidence to justify the trial court’s findings of a breach of implied warranty by defendant; and (3) sufficiency of the evidence to justify the trial court’s finding that defendant was negligent.

Shane’s birth was normal and he developed normally and suffered no unusual reaction from his first inoculation of Quadrigen administered on August 26, 1959. Shane continued to develop normally until the second inoculation which was administered about 5:00 p. m. on October 1, 1959. Immediately thereafter his mother carried Shane from Dr. Graham’s office to her automobile and there discovered that his face was covered with a fine rash. Upon arriving at her home she undressed Shane and noticed a fine reddish rash on the upper part of his body and also a redness at the site of the inoculation. Shortly thereafter Mrs. Stromsodt gave Shane his bottle and he vomited the contents. This was the first time he had done this. Shane then had a “spell,” his eyes rolled back, his head rolled back, he arched his back and he dug his heels into the bed, whined and turned his head from side to side. Mrs. Stromsodt noticed that Shane had a fever and, being concerned, telephoned Dr. Graham. Shane was listless and slept a lot the next two days. He had two more convulsions before November 4, 1959 when he was taken to Dr. Graham for his final shot. On this occasion Quadrigen was not administered but in lieu thereof a triple antigen product was used.

On January 13, 1960, Shane was taken by his family to Dr. Samuel S. Pettit, a physician and surgeon at the Grand Forks Clinic who limited his practice almost exclusively to pediatries. Shane suffered additional and numerous seizures and was later taken by his family to the Mayo Clinic but the Mayo report recites Shane’s condition and prognosis, but made no diagnosis as to the cause of Shane's difficulty. Shane was subsequently observed and examined by Dr. A. R. Natarajan of Minot, North Dakota in St. Joseph’s Hospital from February 4 to February 9, 1966. Dr. Natarajan specializes in neurology and psychiatry. Later Shane was examined on September 25, 1966 at the instance of Parke-Davis by Dr. Blaine McLaughlin, who found that he has serious brain damage, is unable to talk or function normally, has motor difficulty and will be unable to take care of himself. All of Shane’s doctors who examined him agreed with the diagnosis of severe brain damage, motor difficulty, difficulties in speech and that he is handicapped to such a degree that he will never be able to take care of himself and will require close supervision for the rest of his life.

In the early 1940’s American drug manufacturers combined pertussis (whooping cough) vaccine with diphtheria toxoid and tetanus toxoid, thereby making a three-in-one combination known as D.P.T. (Diphtheria-Pertussis-Tetanus). This product has been generally and successfully used by doctors treating young children and is manufactured by a number of drug companies and public health bodies in the United States. Parke-Davis markets its three-in-one product under the name “Trio-gen.”

Diphtheria and tetanus antigens in D.P.T. are toxoids while pertussis antigen is a vaccine. The difference is that toxoids are toxins produced by living *1393 bacteria, separated and treated so as to destroy toxicity while preserving antigenicity. Vaccines are preparations of whole organisms that can be killed or attenuated and suspended in solution. In 1953 Dr. Jonas Salk developed a successful poliomyelitis vaccine, and thereafter Parke-Davis gave consideration to including the poliomyelitis vaccine with its D.P.T. product, Triogen.

Inflammatory reactions involving the central nervous system occasionally follow the injection of various vaccines and serums, including pertussis vaccine, but despite its danger all the medical witnesses agreed that immunization against pertussis was proper and that they would use it or had used it for their own children or for children under their care. Some states even require pertussis vaccinations before a child can be admitted to school. Pertussis vaccine is “biological” as opposed to a drug within the framework of federal statutes and regulations (42 U.S.C.A. § 262 et seq.). Its licensing and regulation are therefore under the Department of Health, Education and Welfare (H.E.W.) rather than through the Food and Drug Administration (F.D.A.). Dr. Margaret Pittman was chief of the Laboratory of Bacterial Products for the Division of Biologies Standards of the National Institutes of Health (D.B.S.) under H.E.W., and reviewed license applications that included pertussis vaccine. Each lot of vaccine had to be specifically released by D.B.S. before it could be placed upon the market. Pertussis vaccine is made of bacteria that has been cultured from bacteria isolated in an actual case of whooping cough and treated to remove its propensity to cause the disease, yet sufficient to cause the formation of antibodies. Thus, a person receiving the vaccine forms antibodies as a reaction against the. introduction of the bacteria and thereby obtains immunity against the disease. The substance in the vaccine that causes one to form antibodies is the “antigen.” The potency test for pertussis vaccine was standardized by D.B.S. in 1953, and thereafter all lots of this vaccine were tested against this standard.. This particular standard remained unchanged until 1961. Additional to the potency test, each lot of pertussis vaccine was given a toxicity test which consisted of injecting massive doses of the vaccine into the bodies of a group of mice and then observing their recovery over a specified period.

Among the many problems in the development of a four-in-one product which included the poliomyelitis vaccine was the selection of an appropriate preservative to protect the final product. Parke-Davis used menthiolate as its preservative in Triogen, but this preservative was not suitable for Quadrigen because it adversely affected the poliomyelitis vaccine, and Parke-Davis ultimately selected benzethonium chloride as its preservative for Quadrigen. Parke-Davis’ trade name for this preservative was Phemerol. There were other technical problems involved in the production of Quadrigen which we do not consider relevant and necessary to discuss.

A product license for production of Quadrigen was granted by H.E.W. on March 25, 1959 and the language of the package insert containing the warning was approved. Quadrigen was first marketed in July, 1959. Additional pertinent facts will be discussed hereafter under the specific issues as required.

We are completely aware of the changing law in this era relating to products liability cases, which has been spawned by life in this modern society. This has resulted in a veritable field day for professors, commentators, practicing attorneys and others, all of which constitutes what might be expressed as an explosion of theories. This also includes treatises, statutes and many judicial opinions as well as a section in the Uniform Commercial Code and the Restatement of Law. 1 Despite all of this, we think it is *1394 not an oversimplification to state that the decision which we reach in this case can be grounded upon settled and elementary rules of law applicable to diversity eases and to the long-standing rule of this court limiting our prerogatives in cases based on factual conclusions when tried to a federal district court sitting without a jury.

In this case the substantive law of North Dakota governs. Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938). The State of North Dakota, as will be seen, has accepted the concept of implied warranty. Our function and limitations in cases tried in the federal district court before a judge sitting without a jury cannot be more plainly or fully articulated than as expressed by the late and eminent Judge John Sanborn in his often repeated words in Cleo Syrup Corp. v. Coca-Cola Co., 139 F.2d 416, 417-418, 150 A.L.R. 1056 (8th Cir. 1943):

“This Court, upon review, will not retry issues of fact or substitute its judgment with respect to such issues for that of the trial court. * * * The power of a trial court to decide doubtful issues of fact is not limited to deciding them correctly. * * * In a nonjury case, this Court may not set aside a finding of fact of a trial court unless there is no substantial evidence to sustain it, unless it is against the clear weight of the evidence, or unless it was induced by an erroneous view of the law. (Citations omitted.)”

Recent cases applying these principles and the “clearly erroneous” rule include Whitson v. Yaffe Iron & Metal Corp., 385 F.2d 168, 169 (8th Cir. 1967); Friedman v. Fordyce Concrete, Inc., 362 F.2d 386 (8th Cir. 1966); Schultz v. State of Nebraska, 353 F.2d 81, 82 (8th Cir. 1965); Consolidated Sun Ray, Inc. v. Oppenstein, 335 F.2d 801, 803-804 (8th Cir. 1964); Pendergrass v. New York Life Ins. Co., 181 F.2d 136, 137 (8th Cir. 1950).

This court is not ready to abandon the traditional and logical rule that cases are not to be tried de novo by a reviewing court, and we see no need in this case to borrow any of the distinguished commentators’ theories on liability in cases under varying fact situations as we do not find it necessary to enlarge or restrict the logical case law of this circuit to reach our conclusion of affirmance of this judgment. We add also that this decision is reached without attaching any significance whatsoever to the testimony of plaintiff’s medical expert, who did not attend the plaintiff and whose evidence Parke-Davis severely assails, for the reason that his expert opinions were based largely on a theory of his own and not on medical certainties or probabilities as is required under North Dakota law, which we will more fully discuss later.

Causation. The burden of proof is “substantive” and under the North Dakota law the burden is upon the plaintiff whether the action sounds in negligence or in breach of warranty. *1395 The question of negligence or probable cause, however, ordinarily is for the trier of the facts. United States Rubber Co. v. Bauer, 319 F.2d 463, 465 (8th Cir. 1963). 2 As hereinbefore mentioned, North Dakota follows the rule that a medical expert is qualified to testify only in terms of medical certainties or medical probabilities and cannot base his opinion on mere medical possibilities. Grenz v. Werre, 129 N.W.2d 681, 689 (N.D.1964); Vaux v. Hamilton, 103 N. W.2d 291, 295 (N.D.1960); Rexall Drug Co. v. Nihill, 276 F.2d 637 (9th Cir. 1960) (applying North Dakota and California law). Nihill is cited with approval by this court in Bauer, supra, 319 F.2d at 465.

We also have noted on numerous occasions that the Supreme Court specifically has not decided in a diversity action whether a state or federal test of the sufficiency of the evidence is to be applied. See United States Rubber Co. v. Bauer, supra, and note the other Eighth Circuit eases, including Twin City Plaza v. Central Surety Co., 409 F.2d 1195, n.8 (8th Cir. 1969); and O’Hare v. Merck & Co., 381 F.2d 286, 288-289 and n. 2 (8th Cir. 1967). A number of the circuits have concluded that the sufficiency question is a federal question. 3 This is not a problem in this case because we have held in Bauer, supra, that the standard of North Dakota is substantially the same as the federal standard.

We turn now to the testimony of the three physicians who attended Shane or who observed and examined him over a period of substantial time.

In determining this issue we are not unmindful that a verdict may not rest on guess or speculation or emerge from “equal probable” causes as articulated in Bauer, supra 319 F.2d at 465. This gives effect to the North Dakota rule heretofore mentioned that a medical expert is qualified to testify only in terms of medical certainties or medical probabilities, and also the rule that the resolution of conflicting testimony, including that of expert witnesses, is for the trier of fact. Baker v. United States, 343 F.2d 222, 224 (8th Cir. 1965). See also Aetna Life Ins. Co. v. Kegley, 389 F.2d 348, 352 (5th Cir. 1967); Roginsky v. Richardson-Merrell, Inc., 378 F.2d 832, 836 (2nd Cir. 1967); Grenz v. Werre, supra; Stokes v. Dailey, 85 N.W.2d 745, 751 (N.D.1957).

Dr. Graham, the family physician, a licensed physician and surgeon and the doctor in whose office the Quadrigen was administered, when asked for his opinion within reasonable medical certainty replied that he could not help but feel that the end result was probably caused by the injection (Quadrigen). 4 Thus, it is noted that Dr. Graham grounds his opinion of causation on probability rather than possibility.

*1396 Dr. Pettit, the pediatrician who attended Shane and also had consultations at the clinic with an associate, Dr. Sil-verman, testified that Shane had evidence of brain damage both on the basis of the history given and the examinations of the child. When asked the direct question of their opinion as to causal connection, Dr. Pettit testified:

“Q. And what was your opinion in consultation? A. That Shane did have evidence of brain damage, both on the basis of the history given and the examination of the child.

“Q. And did you have an opinion, conclusion in consultation as to whether it was causally connected, related to, in reasonable medical expectancy, the inoculation? A. Yes. We did feel that there was a connection, causal, between immunization and brain damage.”

Additionally, on February 13, 1963, when one of plaintiff’s attorneys requested a medical report from Dr. Pettit, he wrote a “To Whom It May Concern” letter and sent a copy or duplicate to Parke-Davis. In this letter Dr. Pettit stated: “It has been our impression throughout our period of observation that Shane’s difficulties are related causally to the second immunization procedure.” Prior to the dispatch of this letter, Dr. Pettit had also discussed this subject with Parke-Davis’ “detail man.”

Dr. Natarajan gave answer to a long, but unobjected to, hypothetical question. This doctor, it will be recalled, had Shane under observation and examination for several days in St. Joseph’s Hospital at Minot from February 4 to February 9, 1966, and he testified that, based on reasonable medical certainty, the competent producing cause of Shane’s condition was the injection of Quadrigen. We quote from Dr. Natarajan’s testimony as follows:

“Now, based upon these facts and upon your own extensive experience and medical knowledge, and from your own examinations conducted over this period of five days at Minot Hospital, do you have an opinion, predicated upon reasonable medical certainty, as to the competent producing cause of the boy’s condition? Will you answer that ‘yes’ or ‘no’ ? A. Yes. (Emphasis supplied.)

“Q. What is that opinion ? A. My opinion is that this boy has suffered an extensive central nervous system damage following the administration of the immunization injections.

“Q. In other words, in your opinion then the condition of the boy and the competent producing cause of that condition is due to the injection of Quad-rigen? A. Of Quadrigen.”

Dr. Natarajan is a Fellow of the Royal College of Physicians of Edinburgh in Scotland, a Fellow in the American Psychiatric Association, a Fellow of the Indian Pyehiatrie Society, a member of the Royal Medical Psychological Association in London and is on the panel of consultants to the Royal Health Organization. He, of course, belongs to the State Medical Society in North Dakota and is licensed there.

Thus, we have the testimony of three expert witnesses, one, the family physician, two, the pediatrician to whom the child was taken by his family within a comparatively short time after his trouble developed, and three, a neurologist who kept Shane under observation for several days, all of whom attribute causation to the injection of Quadrigen. Therefore, unlike Rexall Drug Co. v. Nihill, supra, where the Ninth Circuit, in applying North Dakota law, found from its study of the entire record that “the opinions expressed by appellee’s medical witnesses were in the realm of possibility and not probability,” the opinions here were expressed in the realm of positive probability. We might add that in Nihill the Ninth Circuit said (276 F. 2d at 645):

“Our study of the record has convinced us that the only reasonable inference which can be drawn from the evidence is that the proximate *1397 cause of appellee’s condition is unknown and unproved.”

As opposed to the specific evidence quoted above, the eminent physicians who testified for Parke-Davis adhered in general to the view that it could not be said with reasonable medical certainty what caused Shane’s condition. 5

Faced with the strong evidence here, we have no right to set aside the trial court’s finding of proximate cause.

Breach of Implied Warranty. North Dakota has joined many other states in permitting a recovery for a manufacturer’s breach of implied warranty of fitness for use of its product. Lang v. General Motors Corp., 136 N.W. 2d 805 (N.D.1965). The trial court here held in its Finding No. 10 that:

“The drug Quadrigen administered to the plaintiff was defective and not reasonably fit for the purpose intended. Such defect constitutes a breach of the implied warranty of merchantability.”

Our problem here, therefore, is to determine whether the record before us contains sufficient evidence as would justify the finder of fact in concluding that Quadrigen was defective.

There were many problems connected with the creation of a quadruple antigen vaccine which combined the poliomyelitis vaccine with the well known and successfully used triple combination of D.P.T. In the first place, it was known by all authorities that the preservative used in the triple combination could not be used in Quadrigen because the preservative used in D.P.T. is deleterious to the potency of the poliomyelitis vaccine. As hereinbefore pointed out, Quadrigen (and other D.T.P.P. products) contained a different preservative, benzethonium chloride, which Parke-Davis called Phem-erol. In a paper published in the Quarterly Bulletin of Northwestern University Medical School, Vol. 33, No. 3, p. 259, 1959, Dr. Louis W. Sauer, the author, quoted from the Committee on the Control of Infectious Diseases of the American Acadamy of Pediatrics in 1957 which recommended that if poliomyelitis vaccine is given at the same time as another vaccine, they should be injected separately and at different sites. At the time this article was written, combination preparations of poliomyelitis vaccine and the triple antigen D.P.T. had not been approved and the medical fraternity at that early date indicated that such a combination might soon be approved and supplied commercially, but until then physicians should be cautioned not to mix currently available preparations.

Dr. Margaret Pittman of D.B.S. wrote an article that appeared in the Journal of the American Medical Association of July 7, 1962 entitled “Instability of Pertussis-Vaccine Component in Quadruple Antigen Vaccine.” This article was based in part on a statement previously released by the Massachusetts Department of Public Health to the effect that the potency of the pertussis-vaccine component in the quadruple antigen products containing diphtheria and tetanus toxoids and pertussis and poliomyelitis vaccines was relatively unstable. Dr. Pittman made other such statements in her article as “the stability of potency of pertussis vaccine in single and triple antigen vaccines is in marked contrast to the instability of pertussis vaccine in quadruple antigen vaccines prepared by the same manufacturers.” In her comments, Dr. Pittman stated “the analysis of the results of multiple potency tests of the pertussis-vaccine component in a number *1398 of lots of DTPP has shown that potency of this antigen was not remaining at an acceptable level throughout the dating period of DTPP and that this instability was in contrast to stability in single or multiple antigen vaccines containing diphtheria toxoid and tetanus toxoid.”

In her summary, Dr. Pittman stated “the pertussis-vaccine component in quadruple antigen vaccines containing diphtheria and tetanus toxoids and pertussis and poliomyelitis vaccines has been found to be unstable. This finding is in contrast to potency stability in other products not containing poliomyelitis vaccine * * *. The preservative and the tissue-cell enzymes present in this vaccine may be factors which contribute to instability.” Dr. Pittman’s explanation of the governmental department’s releasing the quadruple vaccine is accounted for by reason of the samples tested having been held in cold storage prior to testing. She stated that “a loss in potency was not readily apparent from the data accumulated by the individual manufacturers who use samples which had been held continuously in cold storage.” She mentioned that in order to provide with reasonable assurance a product containing adequate potency throughout the dating period new provisions and requirements were promulgated which became effective on April 13, 1961 and June 30, 1961. The practical effect of these new requirements on the part of the governmental agency was that a manufacturer could not meet them and products such as Quadrigen had to be withdrawn from the market. Parke-Davis did develop a substitute — a two-vial container used for inoculation which injected the poliomyelitis vaccine separately. Parke-Davis also calls this product Quadrigen.

Prior to the release by the Massachusetts Department of Public Health and Dr. Pittman’s article, the Year Book of Pediatricians, 1959-60 Series, was released. This publication covers more or less the world literature in pediatrics and is a standard reference book for pediatricians, and was referred to by Dr. Pettit in his testimony. He recognized that in this work it is stated that “we have had no experience with the new multiple antigen vaccines, but several personal communications suggest that the reaction rate is considerably higher than that noted with the triple antigen vaccines.”

The eminent doctors on the staff of Parke-Davis were aware of this problem as for example is shown by the question and answer below of Dr. John E. Gajew-ski, Director of Medical Education and Editorial Services for Parke-Davis:

“During the time period which I have previously mentioned, going back within my fencepost, the July ’59 through September of ’61 period, would it be accurate to state that the incidence of elevated fever as reported to you or as coming to your attention was approximately the same for Quadrigen as it was for Triogen? A. The incidence at least as reported appeared to be a little bit more frequent than with Triogen.”

When interrogated on the question of instability in the potency and the variability of it, Dr. McLean stated :

“However, I must say that the evidence here would indicate there was apparently in the overall picture a loss in pertussis potency in the field.”

Plaintiff argues that this defect in the Quadrigen is produced by the preservative’s not having the ability to block the pertussis potency, destroying enzymes which were used in Triogen.

We are not concerned here with what caused the defect in the product, but only whether the product was defective, or, more narrowly, whether there is evidence in this record that would justify the trier of the facts in finding Quadrigen a defective product for its intended use.

Parke-Davis argues vigorously that Quadrigen was not a “defective” product because it is exactly what it was represented to be; that it was within the D.B.S. requirements; that it was pure and uncontaminated and, more importantly, that the pertussis component of Quadrigen was the most likely culprit; *1399 that an association with encephalopathies (inflammation of the brain) in rare instances is an inherent and well-known characteristic of pertussis vaccine; and that one of plaintiff’s doctor’s testimony could not be substantial evidence on which to base the findings because it consisted at best of speculative hypotheses, because his hypotheses were undemonstrated and unknown to experts in the field or in medical literature and because the glaring inconsistencies and misrepresentations in his testimony made his hypotheses unworthy of belief. 6 No contention is made that any one single ingredient of Quadrigen was impure or not exactly what it was represented to be or that the potency tests of the product were not well within D.B.S. requirements. The simple answer to this is that there is no contention that any single ingredient of Quadrigen was anything other than what it was represented to be, or that it was an ingredient different from that represented by the manufacturer, or that any single ingredient was impure or contaminated. The question here is simply a question of whether a combination of the four elements created a defective drug. Obviously, Quadrigen was a defective drug because something about the combination of the poliomyelitis vaccine with the triple antigen product caused the pertussis vaccine to lose its potency. This was known by defendant before the inoculation of plaintiff, although it was undetected at first by the D.B.S. because the reported tests made by the manufacturer, as well as by the D.B.S., were made on the refrigerated sample lots instead of samples being used in the field. In the following discussion on the subject of negligence, it will be shown that Parke-Davis knew or should have known of this defective condition of Quadrigen from the complaints it had from over the country. Additionally, prior to this inoculation it was reported to Parke-Davis by one of its doctor friends in California that a child went into convulsions shortly after injection of Quadrigen and died within twenty-four hours.

All of this adds up to the fact that this record cannot be said to be absent proof that would justify a trier of the facts in holding that Quadrigen was a defective drug. The writer knows of no case yet tried, involving a contested medical issue, where all of the physicians were in total agreement. At least, this occurs in most personal injury cases, and it is for this reason that the jury, or the judge sitting without a jury as the case may be, has the exclusive function of making the conclusive determination provided only that there is substantial medical evidence to justify the findings. This case, where it was so well known by the manufacturer and others that the product under market conditions lost its potency and caused a greater reaction than the triple antigen vaccine, necessarily requires the trier of facts to weigh the evidence and make the factual determination. Beyond that, the trier of facts is not required to close his eyes to the fact that the plaintiff in this ease suffered immediate and severe reaction after the October 1 injection of Quadrigen. So it is that we find from the record evidence that there is ample testimony to justify the trial court’s conclusion that Quadrigen was a defective product for the use intended, and under the substantive law of North Dakota a judgment based on implied warranty is sustainable; and under settled law we, as a reviewing court, do not possess the right to set aside this factual finding.

Negligence. We are also convinced from the totality of the record evidence that the trial court was justified in finding Parke-Davis negligent in its inadequate warning accompanying its product and in finding that Parke-Davis was negligent in testing Quadrigen.

The warning Parke-Davis gave at the time in question was a package insert and a pamphlet for distribution to the medical profession generally reflecting the same type information appearing in the insert. In this case, a Parke-Davis detail man called on Dr. Graham, but he had no more information and imparted no information *1400 beyond that contained in the printed warning. The trial court’s finding mentioned that the insert was in very small print. In its printed warning, Parke-Davis under the headnote “Reactions” stated:

“When given in accordance with suggested methods, local and systemic reactions following the administration of Quadrigen are usually mild. The incidence is usually no greater than is normally experienced with trivalent vaccine.” (Emphasis supplied.)

This is simply not borne out by the evidence in this case which shows that under marketing conditions the pertussis component in Quadrigen lost its potency and caused greater reactions from its inoculation than Triogen. As before stated, Triogen had been a highly successful drug. While some doctors would not use it, yet generally it must be stated that it was successfully used by many practitioners. To tell the practitioner who had successfully used Triogen that reactions from Quadrigen are usually no greater than normally experienced with Triogen simply does not comport with the record facts in this case and the knowledge of the medical experts associated with Parke-Davis and others. In order for a warning to be valid and effective, it must be an adequate warning and warn of known dangers and one that a general practitioner, such as Dr. Graham, could rely upon. Parke-Davis had knowledge that many doctors would not use the product, that other doctors reported to it prior to the inoculation of plaintiff that they were having more severe reactions to the use of Quadrigen than with Triogen, and reports of febrile or higher fever reaction. 7 When this information came to the attention of Parke-Davis, it should have taken steps to warn the medical fraternity of the facts, even though the adverse effect involved only a statistically small percentage of those upon whom it was used. Sterling Drug Co. v. Cornish, 370 F.2d 82 (8th Cir. 1966). It is arguable that the warning above quoted might suffice to put an unusually sophisticated medical man on notice that perhaps the drug should not be used, but certainly it is a matter of common knowledge that there are thousands of rural communities in this country where this type of practice is necessarily restricted to a general practitioner who might be apt to use Quadrigen when being warned merely that the adverse effects thereof are usually no more than with Triogen, when in fact he should have been warned of the reports that Parke-Davis had received (and of which some of the staff doctors admitted having knowledge) that the reaction, at least in many cases, was more violent.

The trial court found that the warning was inadequate in that it suggested that local reactions became more severe “when the child is in the incubative stage of pertussis” and was misleading in implying that only in such instances or in the case of acute pertussis would encephalitic symptoms occasionally occur. The court noted in Finding No. 11 that at the time plaintiff received the injection of Quad-rigen there was no evidence of an epidemic of pertussis nor danger of one in the Grand Forks area, and plaintiff was neither in the incubative nor acute stage of pertussis during any of the times he received the drug.

*1401 Parke-Davis contends that the warning must be considered as a whole and was clear on its face, and additionally that its detail man, Mr. Roche, went over the brochure with Dr. Graham, that Dr. Graham understood it, and that he was not misled. Actually, the detail man just went over the information that was contained in the insert and brochure and there is no record evidence that he was even apprised of the more severe reactions from Quadrigen than from Triogen. Dr. Pettit, the pediatrician who attended plaintiff, testified that there was nothing in the warning that would suffice to put a general practitioner or pediatrician on notice or disclose that Quadrigen has more reaction and is more dangerous than Triogen. 8

We think that, in the light of the information Parke-Davis had and the admission of members of its medical staff that more severe reactions were reported with the use of Quadrigen than with Triogen, Parke-Davis should have given notice of this information to doctors, and that the trial court’s finding from the evidence in this record that the notice did not constitute an adequate warning to the medical profession is properly based on the language of the warning itself and additionally upon the oral testimony of a physician skilled in pediatrics.

The trial court also found that Parke-Davis was negligent in not adequately testing the drug or investigating and properly evaluating the reports of adverse reactions to its use. The record here shows specifically that Parke-Davis was advised by a Los Angeles doctor that a five-month-old baby was given one-half cc. of Quadrigen with development approximately three or four hours later of vomiting and convulsions. About twelve hours after the injection and at the time of the writing of the letter on March 13, 1959, the child had all the signs of pertussis encephalopathy and was quite unresponsive. The doctor further advised that the child did not seem to be getting worse but had not yet begun to improve. This child at four months of age had received an injection of standard D.P.T. without difficulty. The California doctor offered to make any specific test if Parke-Davis would like it done, but he did not receive a reply to his letter from Parke-Davis until March 26, 1959. He then informed Parke-Davis that he was sorry to say that the infant had died the day after his March 13 letter was written. Parke-Davis claimed that the autopsy report showed that it was in no way to blame for the death of this child. This is a factual matter, but even assuming Quadrigen’s lack of causation of death in this particular instance, the vomiting, convulsions, and signs of pertussis encephalopathy experienced subsequent to the injection most certainly should have indicated the propriety of more extensive testing of the product under market conditions and reporting the same to the medical profession.

On July 21, 1958 a doctor from the Los Angeles General Hospital wrote Parke-Davis that Quadrigen had been in use since receiving the supply and was given in the manner outlined in his last letter. He stated that there had been two reactions from the use of Quadrigen and in both cases there had been a soreness at the site of the injection and constitutional symptoms of malaise and fever to 104° F. for forty-eight hours. Neither of these children had had reactions to Triogen and poliomyelitis vaccine given separately in 1957.

The record also contains a letter dated May 25, 1959 from one of the Parke-Davis staff doctors to a Salt Lake City doctor acknowledging that during the past few weeks it had been brought to his attention that a certain lot of Quadrigen had been slightly more reactive than other lots, and stating that if this should be the case with the Quadrigen which the Salt Lake City *1402 doctor was using he would be happy to reclaim it and replace it with a subsequent lot.

On May 27, 1959, an associate physician at the Henry Ford Hospital in Detroit, Michigan wrote Parke-Davis reporting higher fever with Quadrigen than with Triogen. On June 1, 1959, a Detroit doctor wrote Parke-Davis to the effect that there seems to be an unusual amount of pain at the site of injection of Quadrigen for a few minutes after administering. The doctor reported that this was more marked than he had ever noticed in any D.P.T. preparation and certainly greater in contrast to the product of one of Parke-Davis’ competitors. A doctor from Wisconsin wrote Parke-Davis on January 15, 1959 that he had not completed his evaluation of Quadrigen as yet and was gratified with the results, but at this time he would like to request more Quadrigen samples as soon as possible.

All of these letters have a personal ring and undoubtedly and justifiably reflect a friendliness towards the manufacturer. The letters which we mentioned were written prior to the inoculation of plaintiff, and we cannot help but note that subsequent thereto interoffice memoranda at Parke-Davis show that one of its detail men reported in one instance that three of his best pediatricians had stopped using the product because of bad systemic and local reactions. Others reported the loss of two of the best users, reporting that one doctor had used Quadrigen since its introduction and had reported severe temperatures in almost every patient and would not use Quadri-gen in the future. 9 We note also from the record that there is no evidence that Parke-Davis ever reported any adverse reactions to the governmental agency and that it apparently did not treat the adverse reports as of sufficient importance to even have a conference with its staff and seek to find the cause of the trouble. Certainly its physicians were as eminently qualified as those who conducted the tests in Massachusetts which resulted in the report of that agency, and were in a better position than the D.B.S., or anyone else, to determine what was ultimately determined, and that is simply that the addition of the poliomyelitis vaccine to Triogen with the different preservative caused a dilution in the potency of the pertussis antigen. -From the totality of all of this evidence, we think it was for the trier of the facts to determine whether or not its premarket-ing tests were sufficient.

Conclusion. In sum, we are convinced that the issues here are purely factual and the case made difficult only by reason of the short time elapsing between the licensing of the product for market and the inoculation of the plaintiff. We find that the evidence is sufficient to support the conclusions reached by the trial court, and, applying the “clearly erroneous” standard, we are not left with the conviction that a mistake has been committed which would require or permit reversal. United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 92 L.Ed. 746 (1948). This is not to say that if we had tried the case in the first instance we would have found as did the trial court. There is, however, no legal justification for us to set aside this verdict because we might have determined otherwise upon a de novo hearing.

Under the factual situation reflected by this record, the substantive law of North Dakota and our long-established rules as to the functions of a reviewing court in a case tried to the district court sitting without a jury, we must and we do affirm the judgment of the district court.

. See, e.g., “The Liability of the Drug Manufacturer for Drug Reactions” by David W. Louisell, DRUG LIABILITY LITIGATION, Oh. 9, published by The *1394 Institute of Continuing Legal Education (1967); “The Manufacturer’s Duty to Warn of Dangers Involved in Use of a Product,” 1967 Wasli.U.L.Q. 206; “The Manufacture, Testing and Distribution of Harmful New Drugs: The Applicability of Strict Liability,” 28 U.Pitt.L.Rev. 37 (1966-67); “The Fall of the Citadel (Strict Liability to the Consumer)” by William L. Prosser, 50 Minn.L.Rev. 791 (1966), listing states adopting Uniform Commercial Code, § 2-318 (liability of manufacturer or seller); “Products Liability — The Ethical Drug Manufacturer’s Liability” by Paul D. Rlieingold, 18 Rutgers L.Rev. 947 (1964); “The Assault Upon the Citadel (Strict Liability to the Consumer)” by William L. Prosser, 69 Yale L.Rev. 1099 (1960); “Study of Manufacturers’ Liability Without Fault in Tort and Warranty,” 65 Yale L.J. 262 (1955); and “Product Liability: Directions for Use and the Duty to Warn,” 41 Va.L.Rev. 145 (1955). See also Davis v. Wyeth Laboratories, Inc., 399 F.2d 121 (9th Cir. 1968), and the many cases therein cited; Annotation on “Liability of manufacturer or seller for injury caused by drug or medicine sold,” 79 A. L.R.2d 301.

. In United States Rubber Co. v. Bauer, 319 F.2d 463, at 465 (8th Cir. 1963), we stated:

“2. Burden of proof is ‘substantive’ * * *

“3. Under North Dakota law that burden is on the plaintiff, whether the action sounds in negligence, * * * or in breach of warranty * * *

“4. Of course a question of negligence or proximate cause is ordinarily for the trier of fact. * * * (Citations omitted.)”

. See discussion of this question in Den-neny v. Siegel, 407 F.2d 433, 437-443 (3rd Cir. 1969).

. “Q. . . . Within reasonable medical certainty, what was, in your opinion, as treating and caring and yesterday the examining physician, within reasonable medical certainty the cause of the child’s condition? A. Well, I cannot help but feel from the chain of events that took place from the time of the shot up until the present time that, in this case, it is probably the end reaction of the original shot that started the — -that gave the child the convulsion. These convulsions started at that time, they kept right up, according to the mother; and after she had left me, they had to my knowledge, they have still persisted. There has been nothing from the time of the first shot— I mean, the second shot that caused the convulsion. There was nothing in between that I felt could be an additional cause for this, so in my own mind I can’t help but feel that the end result was probably because of the injection.”

. Dr. Russell DeJong testified:

“On the basis of this detailed history that you have given, in spite of the fact that you have given such a detailed history here, I have to say I cannot with reasonable medical certainty state the cause or etiology of this boy’s condition.”

Dr. I. William McLean, Jr., Director of Microbiological Research for Parke-Da-vis, testified:

“X feel that these children [S'hane and Spicer] do have sequelae [any lesion or affection following or caused by an attack of disease], and they certainly had a response in association with an inoculation of Quadrigen.”

. We have given no consideration to this doctor’s testimony.

. Dr. Gajewski of the Parke-Davis staff testified:

“Q. Do you recall whether or not you received during the time period as previously mentioned [July ’59 through September of ’61] reports of temperatures being elevated in excess of 104 degrees? A. Yes.

“Q. Insofar as you know, was any research done or testing done to establish the cause of this range of elevation of temperature? A. All that I recollect that was done, the reports were made to the Biological Control, and investigations perhaps were carried out by them. * S}C iji * ♦

“Q. Do you recall that during the initial year that the Quadrigen in its original form was on the market, that there was some difficulty with maintaining the proper potency in the pertussis antigen component of the vaccine? A. Yes, sir.”

. Dr. Pettit testified:

“Q. Is there anything in there that would warn the ordinarily prudent physician, in your opinion, general practitioner or pediatrician, tliat Quad-rigen has more reaction or is more dangerous than the Triogen? A. No.”

. Cf. Gober v. Revlon, Inc., 317 F.2d 47, 52 (4th Cir. 1963), where the court said:

“We find no error in the Court’s admission of Dr. Fowlke’s testimony that he had treated others subsequently to the plaintiff for dermatitis which he attributed to the defendant’s product. This evidence was not admissible to show notice to the defendant, bat was certainly admissible to show that the defendant’s product adversely affected an appreciable number of persons.”