delivered the opinion of the Court.
In this appeal from a judgment for the plaintiff in an action for personal injuries against a decedent’s estate, we must determine whether the testimony of the surviving plaintiff was corroborated as required by Code § 8.01-397.
A collision occurred on Route 20 in Orange County, on January 1, 1980, between vehicles driven by Marguerite Hereford and by Norman Ray Paytes, Jr. Miss Hereford was killed and Mr. Paytes was seriously injured. Paytes brought an action for personal injuries against the administrator of the Hereford estate, who filed a counterclaim for wrongful death. At the conclusion of a bench trial, the court found for Paytes and dismissed the counterclaim. The administrator appeals, contending that the judgment was based upon Paytes’ uncorroborated testimony.1
The evidence shows that Marguerite Hereford, 19 years of age, accompanied by another young woman, was driving a Datsun south on Route 20 between Unionville and Orange in the early morning hours of January 1, 1980. Norman Paytes, also 19, was driving a Pontiac north on the same section of road. He was alone in the car. The vehicles collided nearly head-on in the approximate center of the road. Route 20 is a two-lane highway at this point, straight and level, and has a pavement width of about 23 feet. The lanes are divided by broken lines. There are shoulders of moderate width outside the pavement on both sides. The weather was clear, the pavement dry, and the visibility unobstructed.
Miss Hereford and her passenger, Suzanne Williams Smith, were killed instantly. Mr. Paytes suffered severe injuries and was removed to a hospital by the rescue squad. There, he was found to have a blood alcohol level of .14 percent two hours after the accident. A witness who qualified as an expert in internal medicine and toxicology testified that, based upon the generally accepted rate at which alcohol is absorbed and eliminated by the body, Paytes’ blood alcohol level at the time of the accident would have been between .16 and .17 percent. This, in the opinion of the witness, would have had a “substantial measurable effect on reaction time, judgment, depth perception and general motor skills, coordination, etc.” Paytes had attended several New Year’s Eve parties.
*607The damage to the vehicles indicated that the right front of each vehicle struck the right front of the other. The point of impact on each vehicle was to the right of center, on the passenger side. The right front wheel of the Datsun was driven back under the front seat. The metal of the Pontiac’s right front fender was bent down onto the tire, immobilizing it. This caused the tire to leave a mark ten to fourteen feet long, which began at or within one foot of the centerline, and which crossed into the southbound lane at a forty-five degree angle to the centerline. The trial court found from the evidence that each vehicle was in the wrong lane of travel:
Now, I make this finding - I find that this collision occurred at a time when the front end of the Paytes vehicle was substantially across the center line with its right wheel either on the center line, or within a foot of it, a foot of it being to the right-hand side in the direction in which he was going. I further find, based upon the physical facts, the physical evidence, the pictures, that the Hereford vehicle was also partially across the center line at the point of impact to a lesser degree but certainly across the center line substantially, that the left front wheel of the Hereford vehicle must necessarily have been substantially across the center line.
Paytes was the only witness to the accident who survived to testify. He stated that he was driving northbound at a speed of fifty-five miles per hour and saw a car coming toward him around a curve ahead, at a high rate of speed, with bright lights on. He said that as it approached him on the straight section of road, it came “way into the middle of the road, square into the middle of the road . . . over the center line. It was more or less half the car was on my side and half the car was on the other side.” He testified that he thought the movement of the oncoming car would bring it further into his lane and that he made a “split second” decision that his best chance of avoiding collision would be to swerve left rather than right. He said that he turned his wheel as far left as he could in the “split second” he had to react, but that he had no chance to remove his foot from the accelerator or to apply the brakes. “As soon as I swerved left the impact took place.”
Code § 8.01-397, as pertinent here, provides that no judgment shall be rendered against the representative of a person inca*608pable of testifying if it is founded upon the uncorroborated testimony of an adverse party. It was enacted to prevent the survivor from having the benefit of his own testimony where, by reason of his adversary’s death, the personal representative has been deprived of the decedent’s version of the facts. Haynes, Executrix v. Glenn, 197 Va. 746, 91 S.E.2d 433 (1956). This statute is a substitute for the common-law disqualification of such a surviving witness for interest.
In Brooks v. Worthington, 206 Va. 352, 143 S.E.2d 841 (1965), we set forth the requirements necessary to fulfill the mandate of the “dead man” statute. Because each case must be decided on its own facts, it is impossible to formulate a fixed rule as to the corroboration necessary in every situation, but the general principles may be stated. It is not necessary that the corroborative evidence be sufficient by itself to support a verdict. “ ‘Confirmation is not necessary for that removes all doubt, while corroboration only gives more strength than was had before.’ ” Id. at 357, 143 S.E.2d at 845 (quoting from Timberlake’s, Adm’r. v. Pugh, 158 Va. at 397, 402, 163 S.E. at 402, 404 (1932)). Corroborating evidence tends to confirm and strengthen the testimony of the witness. It tends to show the probability of its truth. It need not come from other witnesses, but may be furnished by circumstantial evidence. Finally, if the survivor’s testimony is corroborated to some degree, it is unnecessary that it receive corroboration on all material points. Id. at 357, 143 S.E.2d at 845. The corroboration, to be sufficient under the statute, however, must at least tend, “in some degree, of its own strength and independently, to support some essential allegation or issue raised by the pleadings [and] testified to by the [surviving] witness . . . which allegation or issue, if unsupported, would be fatal to the case.” Burton’s Ex’r. v. Manson, 142 Va. 500, 508, 129 S.E. 356, 359 (1925) (emphasis added).
In Penn v. Manns, 221 Va. 88, 267 S.E.2d 126 (1980), we applied the “dead man” statute to the testimony of the survivor in a wrongful death case. Penn, suffering from a gunshot wound, was driven to the hospital by Manns at a high rate of speed. Manns lost control of the car, which left the road and turned over. Penn died as a result of the accident, not from the gunshot wound. When sued by Penn’s administratrix, Manns testified that as he was negotiating a curve on the right, Penn, who was sitting in the front seat beside Manns, said, “Oh, I’m dying,” and fell on Manns’ arm, causing Manns to lose control of the car. Indepen*609dent medical testimony established that Penn’s gunshot wound had penetrated his lung, causing severe bleeding and lung collapse. Expert testimony further affirmed that these conditions might have induced difficulty in breathing, elevated temperature, and, most significantly, fainting. We held that these attendant circumstances furnished sufficient corroboration for Manns’ testimony, and we affirmed a judgment in his favor. 221 Va. at 93, 267 S.E.2d at 130. In Penn, a factual issue upon which the case turned, an allegation “fatal to the case,” received support from evidence which was not dependent upon the credibility of Manns, the survivor.
The case before us is different. If neither party to the collision had been available to testify, the action having been between their personal representatives, then neither representative could have recovered from the other. The evidence of attendant circumstances, in such a case, would have shown that the proximate cause of the collision was each driver’s negligence in driving on the wrong side of the road in violation of Code § 46.1-203. Such a violation would have cast upon each driver the burden of producing evidence in explanation, Fletcher, Adm’r. v. Horn, 197 Va. 317, 89 S.E.2d 89 (1955), but in such a case no explanation would have been available.
The sole feature distinguishing this case from the hypothetical posed above is the testimony of Paytes as to the timing of the departure of the two vehicles from their respective lawful lanes of travel. He testified that before he left the northbound lane, the Hereford car crossed the centerline, entered his lane, and confronted him with a sudden emergency in which the trier of fact might find that he was justified in crossing into the southbound lane. This testimony furnishes Paytes with the explanation required by Fletcher for his driving on the wrong side of the road. But it rests upon Paytes’ credibility alone. It receives no support from the evidence of surrounding circumstances, or from any other source. It presents an allegation which the fact-finder must believe if Paytes is to prevail; therefore its absence would be “fatal to the case.” This is precisely the kind of issue upon which the “dead man” statute wisely requires corroboration.
The trial court, after finding that each car was on the wrong side of the road, as quoted above, stated:
*610Now, of course, that doesn’t resolve the issue. The issue is, why was either vehicle across the center line. Now, it may well have been that Miss Hereford, going down the road saw this vehicle come over to her and she concluded she had better take action to the left-hand side of the road, we don’t know that. If she was here today that may be her explanation. Mr. Paytes is here today and that’s his explanation. Now, his explanation is consistent with certain physical facts. I don’t mind saying that had she come in with the same explanation that that would have been consistent with the physical facts but we don’t have that explanation, we do have his explanation. Now, can we say the explanation given by Paytes, that is consistent with the physical facts, is to be rejected, I don’t think we can. I don’t think we can reject his testimony given here today in a forthright and forward manner judging him by his appearance on the witness stand, his manner while he was testifying, his behavior while testifying,
This statement makes it clear that the credibility of Paytes, unsupported by the ambivalent circumstantial evidence, was the sole factor determining his recovery.2 It confirms our conclusion that this case represents the very situation the statute was designed to prevent: an opportunity for the survivor to prevail by relying on his own unsupported credibility, while his opponent, who alone might have contradicted him, is silenced by death.
For the reasons stated, the judgment appealed from will be reversed and final judgment entered here for the appellant.
Reversed and final judgment.
No appeal was taken as to the court’s ruling on the counterclaim.
Although an explanation of the fact-finder’s reasoning would have been lacking if the case had been decided by a jury’s general verdict, the result here would necessarily be the same. It is dictated not by the reasons expressed for the court’s ruling, but by the failure of corroboration of the surviving party’s testimony on any issue “fatal to the case.”