SOBELOFF, Circuit Judge:

This action was brought under the Federal Tort Claims Act, 28 U.S.C. § 1346, to recover damages for the death of Carol Greitens. The plaintiff, administrator of her estate, alleges that death was due to the negligence of the doctor on duty at the dispensary of the United States Naval Amphibious Base, Little Creek, Virginia, in diagnosing and treating her illness. The District Court, concluding that the evidence was insufficient to establish that the doctor was negligent, or that his eoncededly erroneous diagnosis and treatment was the proximate cause of her death, dismissed the complaint. From this action, the administrator appeals.

The decedent, 25 years of age, had been a diabetic since the age of 13, although the condition was under control. As the wife of a Navy enlisted man, she was entitled to medical care at the dispensary. Mrs. Greitens’ husband brought her to the dispensary at about 4 a. m. on August 25, 1963, suffering from intense abdominal pain and continual vomiting which had begun suddenly an hour before. The corpsman on duty in the examining room procured her medical records, obtained a brief history, took her blood pressure, pulse, temperature, and respiration and summoned the doctor on duty, then asleep in his room at the dispensary. The doctor arrived 15 or 20 minutes later and after questioning the patient concerning her symptoms, felt her abdomen and listened to her bowel sounds with the aid of a stethoscope. Recording his diagnosis on the chart as gastroenteritis, he told Mrs. Greitens that she had a “bug” in her stomach, prescribed some drugs for the relief of pain, and released her with instructions to return in eight hours. The examination took approximately ten minutes.

*629 The patient returned to her home, and after another episode of vomiting, took the prescribed medicine and lay down. At about noon, she arose and drank a glass of water, vomited immediately thereafter and fell to the floor unconscious. She was rushed to the dispensary, but efforts to revive her were unsuccessful. She was pronounced dead at 12:48 p. m. and an autopsy revealed that she had a high obstruction, diagnosed formally as an abnormal congenital peritoneal hiatus with internal herniation into this malformation of some of the loops of the small intestine. Death was due to a massive hemorrhagic infarction of the intestine resulting from its strangulation.

The plaintiff contends that the doctor at the dispensary did not meet the requisite standard of care and skill demanded of him by the law of Virginia. Compliance with this standard, the plaintiff maintains, would have required a more extended examination and immediate hospitalization. More specifically, plaintiff’s expert witnesses, two general practitioners in the Norfolk-Virginia Beach area, testified that, according to prevailing practice in the community, the doctor should have inquired whether the patient had had diarrhea and should have made a rectal examination to determine whether the patient was suffering from an obstruction rather than from gastroenteritis. While the latter condition does not ordinarily require immediate radical treatment, a high obstruction is almost invariably lethal unless promptly operated upon. Plaintiff’s experts further testified that on observing the symptoms manifested by Mrs. Greitens, the procedure of general practitioners in the community would have been to order immediate hospitalization. This the dispensary physician failed to do, although the Naval Hospital in Portsmouth was available to him.

The standard of care which Virginia law exacts from a physician, in this case a general practitioner, is stated in Reed v. Church, 175 Va. 284, 8 S.E.2d 285, 288 (1940), as follows:

A physician holds himself out as possessing the knowledge and ability necessary to the effective practice of medicine * * *. However, he is not an insurer, nor is he held to the highest degree of care known to his profession * * *. He must exhibit only that degree of skill and diligence employed by the ordinary, prudent practitioner in his field and community, or in similar communities, at the time.

Accord, Alexander v. Hill, 174 Va. 248, 6 S.E.2d 661 (1940); Fox v. Mason, 139 Va. 667, 124 S.E. 405 (1924). See Shepherd, The Law of Medical Malpractice in Virginia, 21 Wash. & Lee L.Rev. 212 (1964). Thus, if he uses ordinary care in reaching his diagnosis, and thereafter acts upon it, he incurs no liability, even if the diagnosis proves to be a mistake in judgment.

It is undisputed that the symptoms of high obstruction and of gastroenteritis are quite similar. The District Court placed great emphasis on this fact as an indication that the doctor’s erroneous diagnosis was not negligent, but was merely an error of judgment. It would seem, however, that where the symptoms are consistent with either of two possible conditions, one lethal if not attended to promptly, due care demands that a doctor do more than make a cursory examination and then release the patient. See Jenkins v. Charleston Gen. Hospital & Training School, 90 W.Va. 230, 110 S.E. 560, 22 A.L.R. 323 (1922), holding that where a “partial and very hurried investigation” was made, the physician was liable for failure of his diagnosis to disclose an injury which caused detriment to the patient. The fact that an intestinal obstruction is a rare occurrence, and that some form of gastroenteritis is the more likely of the two conditions, *630 does not excuse the failure to make inquiries and perform recognized additional tests that might have served to distinguish the one condition from the other. The dispensary doctor himself, as well as the experts for both sides, agreed that an inquiry as to diarrhea and a rectal examination were the “proper procedure” and “the accepted standard” in order to be able to rule out gastroenteritis and to make a definite diagnosis of high intestinal obstruction. If he had made the inquiry which he admits was the accepted standard, he would at least have been alerted to the fact that the case was one calling for close observation with a view to immediate surgical intervention if the graver diagnosis were confirmed. In these circumstances, failure to make this investigation constitutes a lack of due care on the part of the physician. It was stated in Kelly v. Carroll, 36 Wash.2d 482, 494, 219 P.2d 79, 86, 19 A.L.R.2d 1174 (1950), cert. denied, 340 U.S. 892, 71 S.Ct. 208, 95 L.Ed. 646 (1950), a case in which an erroneous diagnosis had led to improper treatment, that “if there was a possibility that it was appendicitis, he [defendant] had no right to gamble with [decedent’s] life, on the theory that it might be something else.” Only if a patient is adequately examined, is there no liability for an erroneous diagnosis. 1

Our conclusion that the physician was negligent in his diagnosis and treatment of the patient is not inconsistent with Fed.R.Civ.P. 52(a), which declares that the trial judge’s findings of fact are not to be disturbed unless clearly erroneous. This Rule comes into' play primarily where the trial judge as. fact finder has had to reconcile conflicting testimony. Where the veracity of witnesses is in issue, the decision is for the judge who has had the opportunity to see and evaluate the witnesses’ demeanor. The trial court’s findings of fact on conflicting evidence will not be disturbed by the appellate court unless, clearly erroneous. United States v. General Motors Corp., 384 U.S. 127, 141 n. 16, 86 S.Ct. 1321, 16 L.Ed.2d 415. (1966); Walling v. Gen. Industries Co.,. 330 U.S. 545, 550, 67 S.Ct. 883, 91 L.Ed. 1088 (1947); Nationwide Mutual Ins. Co. v. DeLoach, 262 F.2d 775, 778 (4th Cir. 1959). But we are dealing here with the testimony of expert witnesses who are not in controversy as to the basic facts; thus, the opportunity of the trial court to observe the witnesses is of limited significance. It has often been held that, where the trial court’s conclusions are. *631 based on undisputed facts, they are not entitled to the finality customarily accorded basic factual findings under Rule 52(a). United States v. General Motors Corp., supra; United States v. Parke, Davis & Co., 362 U.S. 29, 44, 80 S.Ct. 503, 4 L.Ed.2d 505 (1960); Shapiro, Bernstein & Co. v. H. L. Green Co., 316 F.2d 304, 306-307 (2d Cir. 1963) ; St. Louis Union Trust Co. v. Finnegan, 197 F.2d 565, 568 (8th Cir. 1952).

The question before us is not one of fact in the usual sense, but rather whether the undisputed facts manifest negligence. Although the absence of a factual dispute does not always mean that the conclusion is a question of law, it becomes so here since the ultimate conclusion to be drawn from the basic facts, i. e., the existence or absence of negligence, is actually a question of law. For this reason, the general rule has been that when a judge sitting without a jury makes a determination of negligence, his conclusion, as distinguished from the evidentiary findings leading to it, is freely reviewable on appeal. Mamiye Bros. v. Barber Steamship Lines, Inc., 360 F.2d 774, 776 (2d Cir. 1966). The determination of negligence involves not only the formulation of the legal standard, but more particularly in this case, its application to the evidentiary facts as established; and since these are uncontested, there is no basis for applying the “clearly erroneous” rule. United States v. Parke, Davis & Co., supra; Kippen v. American Automatic Typewriter Co., 324 F.2d 742, 745 (9th Cir. 1963) ; Romero v. Garcia & Diaz, Inc., 286 F.2d 347, 355 (2d Cir. 1960), cert. denied, 365 U.S. 869, 81 S.Ct. 905, 5 L.Ed.2d 860 (1961); Galena Oaks Corp. v. Scofield, 218 F.2d 217, 219 (5th Cir. 1954).

The government’s expert opined that the dispensary physician exercised “average judgment,” but analysis of his entire testimony points unavoidably to the opposite conclusion. Revealing are his statements that it was wrong not to inquire about diarrhea, conceding that that is one question that one usually asks,” and that given a patient with abdominal pain of one hour’s duration, it is too soon “to expect anybody to come up with a proper diagnosis.” Furthermore, his opinion was predicated upon a factual assumption not permissible in this case. His assumption was that the dispensary physician had made only a “working” or “tentative” diagnosis, which the expert felt to be appropriate in view of the fact that the symptoms had their onset such a short time before. However, the uneontradicted evidence indicates that this was not a “tentative” diagnosis. U

The examining doctor himself testified that he had already considered and ruled out at the beginning of his examination the possibility of an obstruction without making the additional differentiating diagnostic tests. He said that his only reason for asking the patient to return eight hours later was because her diabetic condition could become complicated by a case of gastroenteritis. A further indication of the final nature of the diagnosis is his notation of gastroenteritis on the chart, made without further qualification. He also testified that he told the woman not to return for eight hours, regardless of the persistence of pain; yet even the government’s expert testified that if abdominal pain were present for “three or four hours and wouldn’t go away, you would probably have to operate.” By releasing the patient, the dispensary physician made his diagnosis final, allowing no further opportunity for revision, and this prematurely determined final diagnosis was based on an investigation not even minimally adequate.

On careful scrutiny, therefore, the government’s expert is seen to have demonstrated that the examiner did not conform to the required standard of care. Coupled with the explicit testimony of the plaintiff’s experts, the government’s testimony leads us inevitably to the conclusion that the doctor was negli *632 gent as a matter of law. We think that the District Court gave undue weight to the purely conclusory opinion of the government witness. The District Court is not bound by his statement that “average judgment” had been exercised, nor are we bound by it. Only the standard of care is to be established by the testimony of experts. If under the undisputed facts the defendant failed to meet that standard it is not for the expert but for the court to decide whether there was negligence.

The government further contends that even if negligence is established, there was no proof that the erroneous diagnosis and treatment was the proximate cause of the death, asserting that even if surgery had been performed immediately, it is mere speculation to say that it would have been successful. The government’s contention, however, is unsupported by the record. Both of plaintiff’s experts testified categorically that if operated on promptly, Mrs. Greitens would have survived, and this is nowhere contradicted by the government expert. Price v. Neyland, 115 U.S.App.D.C. 355, 320 F.2d 674, 99 A.L.R.2d 1391 (1963), decided under Virginia law, held that a doctor was liable for negligent diagnosis, although even when correctly diagnosed, the disease requires immediate treatment for success.

When a defendant’s negligent action or inaction has effectively terminated a person’s chance of survival, it does not lie in the defendant’s mouth to raise conjectures as to the measure of the •chances that he has put beyond the possibility of realization. If there was any substantial possibility of survival and the defendant has destroyed it, he is answerable. Rarely is it possible to demonstrate to an absolute certainty what would have happened in circumstances that the wrongdoer did not allow to come to pass. The law does not in the existing circumstances require the plaintiff to show to a certainty that the patient would have lived had she been hospitalized and operated on promptly. Harvey v. Silber, 300 Mich. 510, 2 N.W.2d 483 (1942). 2

An apt analogy is found in a case which arose in this circuit. 3 In Gardner v. National Bulk Carriers, Inc., 310 F.2d 284, 91 A.L.R.2d 1023 (4th Cir. 1962), cert. denied, 372 U.S. 913, 83 S.Ct. 728, 9 L.Ed.2d 721 (1963), a seaman, last *633 seen on board at 6 p. m., was not discovered to be missing until 11:30 p. pi, A search was then made of the ship, but the master did not alter the speed or course of the vessel or take other steps to find the missing seaman. It was contended that there was no causal connection between the failure to reverse the course of the ship to conduct a search and the seaman’s death, since there was no definite showing that when it was discovered that he was missing, he was still alive and could have been saved. This court held, however, that the master’s duty to attempt a rescue

is of such a nature that its omission will contribute to cause the seaman’s .death. The duty arises when there is a reasonable possibility of rescue. Proximate cause is tested by the same standard, i. e., causation is proved if the master’s omission destroys the reasonable possibility of rescue. [Citation omitted.] Therefore, proximate cause here is implicit in the breach of duty. Indeed, the duty would be empty if it did not itself embrace the loss as a consequence of its breach. Once the evidence sustains the reasonable possibility of rescue, ample or narrow, according to the circumstances, total disregard of the duty, refusal to make even a try, as was the case here, imposes liability. Id. at 287. [Emphasis supplied.]

In sum, the dispensary physician’s negligence in failing to make a thorough examination and in omitting standard diagnostic tests, led to an erroneous diagnosis. Because of this, he sent the patient home with instructions not to return for eight hours, rather than immediately admitting her to a hospital. •Since the uneontradicted testimony was that with prompt surgery she would have survived, the conclusion follows that the dispensary doctor’s negligence nullified whatever chance of recovery she might have had and was the proximate cause of the death.

Judgment reversed and cause remanded for the determination of damages.

Notes

. Numerous cases have held that a physician has a duty to make proper use of all available diagnostic aids to establish a firm basis for the diagnosis and choice of treatment. See, e. g., Price v. Neyland, 115 U.S.App.D.C. 355, 320 F.2d 674, 99 A.L.R.2d 1391 (1963) (pediatrician negligent in stopping short of making all possible diagnostic tests which might have enabled him to distinguish physiologic from pathologic jaundice); Kingston v. McGrath, 232 F.2d 495, 54 A.L.R.2d 267 (9th Cir. 1956) (ordinary skill and care required further examination of patient and taking of additional X-ray pictures even though first X-rays did not disclose fracture); Dowell v. Mossberg, 226 Or. 173, 355 P.2d 624 (1960), rev’d on other grounds on rehearing, 359 P.2d 541 (1961) (failure to diagnose disease as diabetes negligent because blood sugar test not performed); Harvey v. Silber, 300 Mich. 510, 2 N.W.2d 483 (1942) (doctor’s erroneous diagnosis of position of bullet in decedent’s body negligent because of reliance on manual examination rather than X-rays); Peterson v. Hunt, 197 Wash. 255, 84 P.2d 999 (1938) (diagnosis of ovarian cyst as pregnancy negligent because of failure to employ standard “rabbit test”); Ramberg v. Morgan, 209 Iowa 474, 218 N.W. 492 (1929) (erroneous diagnosis of auto accident victim’s condition as intoxication, rather than skull fracture, negligent because of failure to give more thorough examination, including X-rays); Coleman v. Wilson, 85 N.J.L. 203, 88 A. 1059 (1913) (negligent failure to analyze tissue of growth in decedent’s nostril, leading to improper operation, when analysis would have revealed non-malignant character of growth); Note, Problems of Negligent Malpractice, 26 Va.L.Rev. 919, 922 (1940).

. In Harvey, the defendant negligently diagnosed the decedent’s wound as merely superficial, whereas a correct diagnosis would have indicated that the intestine had been pierced, and that an operation would be necessary to stop the hemorrhaging. As here, the expert witnesses in that case agreed that the decedent could not have survived without an operation. Since the negligent diagnosis was the proximate cause of the failure to operate, and there was testimony to the effect that there was a probability that an operation would have saved decedent’s life, it was held that the negligent diagnosis was the proximate cause of death. The appellate court found that the jury had been properly charged when told that it was “not incumbent on the plaintiff to show that to a certainty surgical intervention would have saved his life.” Id. 300 Mich, at 521, 2 N.W. 2d at 488.

. Other circuits have similarly held that if the victim might have been saved by a precaution which the defendant negligently omitted, the omission is deemed to have caused the harm, even though it is not possible to demonstrate conclusively that the precaution would in fact have saved the victim. See, e. g., Kirineich v. Standard Dredging Co., 132 F.2d 163 (3d Cir. 1940) (seaman who could not swim was thrown inch heaving line rather than larger and more buoyant object) ; Zinnel v. United States Shipping Board, 10 F.2d 47 (2d Cir. 1925) (whether guard rope, absence of which constituted negligence of defendant, would have prevented plaintiff’s intestate from being washed overboard). In the latter case, it was stated that although nobody could be sure intestate would have seized rope or that it would have stopped his body, the court was not “justified, where certainty is impossible, in insisting upon it.”