Hall, Judge.
The question of first importance, upon which depends the necessity to review all the others, is whether or not there is any evidence sufficient in law to- support the jury’s verdict that plaintiff’s husband, Ercell B. Miller, died before midnight December 19, 1959. The record contains no positive evidence supporting this conclusion. There is supporting circumstantial evidence, in the testimony of witnesses set out below. (We exclude from consideration for the moment the evidence opposing the conclusion).
The evidence showed that between 9 p.m., December 19, and 10 a.m., December 20, the temperature ranged from 34° to 43° F. C. C. Tays, the owner of the land where Miller’s body was found, testified that he discovered Miller’s body on the morning of December 20’, 1959, as he was coming back home after taking his wife to Sunday School, where she had to be before 10, and after “fooling around” on his property at least 30 minutes. “. . . And I just walked up on the bank of the little old creek and looked down there, and there lay that guy just flat on his back there with his hands up about like this, right in the creek right behind this car. . . The body that I first saw—its head and part of its shoulders and its hands was out of the water. He was laying just about like that and—both hands kind of extended. And his face, head and shoulders, nose and mouth then, were above the water and part of his shoulders or maybe all. I was standing right there when they pulled them out . . . I let them help him pull them out, and he was stiff. He was stiff all over—just about like handling a plank. I’m speaking of the first body. The other body was stiff, too. . . It—he was laying—the deepest part of the water was where his feet was, you see, and part of his shoulders and his head and neck and his hands was above the water.”
Bill Mendenhall, a highway patrolman: “I found this Mr. Miller behind the automobile laying partially in the water and partially out of the water. . . It was 10:35 a.m. when I received this call. . . The stream of water Was about five or [557]*557six feet wide. The depth varied in places, but there where the car was, it was about a foot. Miller was partly in and partly out of the water. His arms was out and—one of his arms was out of the water and part of his head was out of the water and the upper part of his chest was out of the water. The water wasn’t deep enough for his whole body to lay down in this water. I don’t remember just how his face was in that water. I cannot remember whether his nose and mouth were under the water or not. . . The entire bodies seemed to- be' stiff enough to indicate that they Were dead bodies to me. They looked like they were stiff, that is, legs, arms, feet and the bodies all over seemed to be stiff. I helped put Miller in the ambulance. I do not recall in lifting him whether he was limber or stiff. I saw them at Hillcrest Hospital when the bodies were disrobed and after they were disrobed. They undressed Miller while I was there, but Moore was undressed when I arrived. That’s the only occasion I’ve had to see a stiff person. Now, how stiff he was, I don’t know. . .”
Ray Bachus, a county investigator: “. . . As to his coloring or complexion, at that time he was—he was gassed up. In other words, he was blue at that time, at the time that we arrived there. . . He didn’t present all of the same appearance he presented in life, as I guess no dead body does; but after a while, he was gassed up and blue. By gassed up, I mean there was some swelling. Rigor mortis had set up, in other words. . . His stomach had more of a swelling than his chest or aims. He was of a darkish bluish color. Not all over, his feet and arms, they’re just more or less—not the—-they’ll just be spots over the body. Rigor mortis set up through stiffness of the joints. There were spots on him. His stomach was swollen and I could tell from the appearance that he Was stiffened. It was between nine and ten when I arrived at the hospital. I would say that it would take—something like 20 to 30 minutes to get him there. That would then have fixed it so that the call must have been by 9:30 or before. . . Well, if rigor mortis isn’t set up in a body, why; you can tell. But it had set up in him. In other words, it would have made it difficult, unusually, for me as a layman to determine. I depended on [558]*558the intern. I had confidence in the intern. . . I just accepted whatever they gave me at that time and took his word for it. . . I got there a little before ten and saw him blue and swollen, with rigor mortis set in, so I knew good and well that he died a good little while before ten. . . When I saw him swollen and blue, I knew he had been dead for some length of time before that.”
Clarence Foster, an embalmer: . . Rigor mortis is the muscular juices leaving the joints of a person’s body which causes stiffness and can set in anywhere from one minute to twenty-four hours after death. It remains for a spell. Like I said, it will set in and it will take it anywhere from one minute to twenty-four hours to set in and pass off. It passes off and the person is not stiff anymore. Warm weather would cause it to set in and pass off more quickly it has been my experience to observe than in cool weather. Well, as a general rule it starts in the joints first and finally permeates the entire body. If 1 assume a situation where the weather is somewhere between 40 degrees and freezing but maybe 35 degrees, cold enough for frost but not cold enough for ice, would that or not cause rigor mortis to set in more slowly than on a warm day like 65 or 70 degrees, in my opinion? It would set in more slowly in cool weather than it would in warm weather. If a person say thirty years old who was about five feet and nine inches tall weighed about 130 pounds, was found in water nearly up to his shoulders in water his head and part of his aims and shoulders out, the lower part of his body in water but not freezing water maybe 40 degrees, 36-7 or 8 or 35 degrees and if that person were found in that condition, deceased and if he were stiff as a plank or stiff as a board, each and every part of his body was stiff, his head his arms and hands and fingers, toes, rigor mortis had set in completely in that kind of weather, in my opinion. I would say that it would take several hours for rigor mortis to1 actually set in and start to work. If it is finally completed though; if it was rigor mortis all over the body, in every part of it, I would say that would have been as much as 12 hours. If I assume, in addition, that there was an extended protuberance of the abdomen, swollen, and there were blue splotches or dark splotches about [559]*559over the body, that would indicate, in my opinion, that the body had been deceased for a great length of time due to this fact that what would cause the swelling and the splotches would be that after death the body fluids have a tendency to seek their lowest level and after it does that, then what we call tissue gas sets up which causes a swelling and enlargement of the abdominal area. . . It would be impossible to determine with any degree of accuracy when the man died, but from all indications it was quite a while. From the indications you were speaking of a small man weighing around 130 or 135 pounds, temperature around 33 and 35 and 40, and part of him being in water, too, would have a tendency to slow the tissue gases from working as fast as it normally would if it was in weather such as we are having here today. It could possibly have slowed them so that it took over twenty-four hours.”
Carl Mock, an embalmer and coroner: “. . . When a person dies and rigor mortis sets in, that’s a rigid stiffening of the body. The hotter temperatures cause the rigor mortis to set in the more quickly, ordinarily and generally. In the colder temperature, everything would be more dormant. If I assume this to be true, that a person 30 years old, about 5 feet 9 inches-tall, who weighed about 130 pounds, was found, with most of his body in water and the upper part, shoulders, head, and face out of the water, maybe part of his arms, and the water was somewhere between 40 degrees and maybe 35 or 36 degrees, not freezing, but he was found there in that water and when he was taken out of the water, it was observed that he was completely stiff, as the witness said, ‘stiff as a plank’ or ‘stiff as a board’ in every part of him, head, feet, hands all of the limbs and all of the appendages were stiff. Now, that being the case and assuming that was a true situation, I could give you an opinion as to how long such a person had been dead. I wouldn’t want to say to the hour but—I would say it would be a considerable time. In my opinion as to rigor mortis, it can start immediately after death or it can be delayed from 8, 10 or 12 hours, or even sometimes longer. Now, my definition is that as rigor mortis sets on, it starts at the head to the neck, shoulders, and goes out. It goes to the feet last. It goes off the same [560]*560way it came on. It goes out the head first. And, as I said, at 10 or 12 hours, that is the onset but actually it takes sometimes according to the temperature how long that will complete itself. Not committing myself to testifying to a fact, just in my opinion, if these things were true, such a person like that likely would have been deceased for 12 hours or more. If I assumed that there were spots over the body of a darkish bluish color and a cut over the eye, spots over the body of a darkish, bluish color and a swollen abdomen, in addition to these other things, that would help to. fortify my opinion as to the length of time. With that addition, if I had any doubts, I would assume possibly then that it was that long or longer.”
We shall apply to this testimony the law declared in previous decisions. Ga. Ry. &c. Co. v. Harris, 1 Ga. App. 714 (57 SE 1076) a leading case often cited, held: “. . . testimony must be such as to reasonably establish the theory relied upon, and to preponderate to that theory, rather than to any other reasonable hypothesis, (a) . . . the sufficiency of the evidence is for the jury, yet . . . the circumstances shown must in some appreciable degree tend to. establish the conclusion claimed.” “When the party upon whom the burden of an issue rests seeks to carry it, not by direct proof, but by inferences, he has not . . . submitted any evidence for a jury’s decision, until the circumstances he places in proof tend in some proximate degree to establish the conclusion he claims; and for this, the facts shown must not only reasonably support that conclusion, but also, render less probable all inconsistent conclusions . . by a mere preponderance. . .” P. 717.
In Carroll v. Atlanta Paper Co., 7 Ga. App. 584, 586 (67 SE 680), Judge Powell, regretfully affirming a nonsuit for the defendant, stated: "As she was the plaintiff, and as she was unaided in her proof by any statutory presumption, she had the burden of proving the negligence of the master; and as proof by circumstantial evidence works according to the process of exclusion, she had the burden of excluding as being unreasonable, or as being not so reasonable as the theory of the defendant’s negligence, the theories of accident and of the events having-occurred through the operation of causes as to which she had assumed the risk.”
[561]*561The theory of the burden to exclude other reasonable hypotheses is explained in Radcliffe v. Maddox, 45 Ga. App. 676, 682, 683 (165 SE 841): “It is for the jury to say whether or not they preponderate to' that theory as against all other reasonable, but less probable, hypotheses. . . Where the circumstances shown do, as a matter of law, in some appreciable degree tend to establish the hypothesis claimed, and in the minds of the jury preponderate to tihat hypothesis rather than to any other reasonable hypothesis, this is the equivalent of excluding all such other less probable hypotheses. In other words, the jury could not find that circumstantial evidence of real probative value preponderated in favor of one theory as against all other reasonable but less probable theories, without excluding the theories thus rejected. In neither criminal nor civil cases is it required that the proved circumstances shall show consistency with the hypothesis claimed and inconsistency with all other reasonable theories to the point of logical demonstration. In civil cases all other reasonable theories are excluded when proved circumstances of real and actual probative value cause the jury to find that the preponderance of the evidence is in favor of the hypothesis claimed, as against all other reasonable but less probable theories. Where a decision is required between two or more antagonistic theories, an authorized finding that the evidence preponderates to one theory as against all the others necessarily carries with it a finding that the rejected theories are excluded.”
In Dodd v. Callaway, 76 Ga. App. 629, 636 (46 SE2d 740), a directed verdict for the defendant was reversed. The court stated: “Under the principles of law set out in the cases cited herein, we think that this case should have been submitted to' the jury. Whether the evidence preponderated in favor of the theory that a train of the defendant inflicted the mortal wounds upon the deceased, as contended by the plaintiffs, rather than to some other reasonable theory, and whether the defendant was negligent in one or more of the ways claimed by the plaintiffs, and whether such negligence was the proximate cause of the death of the deceased, are questions which we think should have been submitted to the jury.”
These facts appear from the evidence: When patrolman Men[562]*562denhall arrived at the scene, shortly after 10:35 a.m., December 20, 1959, according to his testimony, Miller’s body appeared to be stiff, “that is legs, arms, feet and the bodies all over seemed to be stiff.” When county investigator Bachus saw the body at the hospital it was “gassed up and blue” in spots, with some swelling, and looked stiffened. This was between 9 and 10 o’clock on December 20, according to his testimony. It would take 20 or 30 minutes to move the body from the scene of the accident to the hospital. The expert opinion of two witnesses based upon these facts was that the body had been deceased for 12 hours, or more than 12 hours. Even if the jury believed Mendenhall rather than Bachus as to the time when the body appeared in the condition described, the time of death according to the expert opinion would be before midnight. The above facts and the opinions based thereon supported the verdict of the jury.
To pass on the denial of defendant’s motion for directed verdict, we must review the evidence to determine whether or not the verdict is contrary to the evidence or whether the evidence demanded a verdict for the defendant. Echols v. Thompson, 211 Ga. 299, 300 (84 SE2d 423).
There were two witnesses who testified that they saw Miller alive after midnight, December 19th. The plaintiff contends that each of these witnesses was impeached. If either of them is unimpeached, his testimony must prevail over the circumstantial evidence favorable to the plaintiff. Frazier v. Ga. R. & Bkg. Co., 108 Ga. 807 (33 SE 996); Myers v. Phillips, 197 Ga. 536, 542 (29 SE2d 700).
William Vaught rented a room at the same house as did the deceased Miller. Pie testified that he saw Miller after midnight, and later overheard a conversation between Miller and Moore (who was killed in the same accident) in which one said to the other that it was 15 minutes until 3; and that he (Vaught) looked at his clock and confirmed this time. On cross-examination this witness testified that the weather the day of the fatal accident was cold but not freezing. The temperature was relevant to the case as affecting, according to expert opinion in evidence, when after death rigor mortis of the deceased’s body would have set in. Plaintiff sought to impeach the witness by [563]*563proving an inconsistent statement made by him the day before the deposition in the manner prescribed by Code § 38-1803. Plaintiff’s counsel asked Mr. Vaught this question: “And then didn’t I ask you as to whether it was quite cold, and then didn’t you proceed to tell me that it wasn’t, and made a comparison with yesterday’s weather? Isn’t that so?” The witness answered: “Yes. But what I had in mind was not the temperature but the fairness of the weather.” Randall Evans, Jr., plaintiff’s counsel, then testified: “. . . I said, ‘It was a kind of a cool day, wasn’t it?’ So he [Vaught] said, ‘Oh, no, it wasn’t cool at all.’ He said it was about like today. I sa-id, ‘Now, that was on December 19th or 20th, I am talking about, and this is in August.’ He said, ‘Our winter doesn’t set in until late.’ I believe he said the last part of January or the first of February, I am not sure, but way later than that. He said, ‘We have quite unusual weather.’ He said, ‘I would compare it with today [in August].’ ”
Whether or not a witness -has been successfully impeached is a question for the jury. Lewis v. State, 91 Ga. 168 (16 SE 986); Williams v. State, 69 Ga. 11, 14; Powell v. State, 101 Ga. 9, 10 (29 SE 309, 65 ASR 277); Huff v. State, 104 Ga. 521 (30 SE 808); Blumberg v. Grant, 34 Ga. App. 253 (129 SE 144); Wilson v. Gray, 34 Ga. App. 320 (129 SE 297); Metropolitan Life Ins. Co. v. Joye, 77 Ga. App. 357 (48 SE2d 751). When evidence is introduced in an effort to impeach a witness, the jury can judge him unworthy of belief whether the inconsistency shown is due to mistake or wilfulness on the part of the witness. Blumberg v. Grant, supra; Green, The Georgia Law of Evidence 335, § 135.
The proof of Vaught’s previous inconsistent statement authorized the jury to disbelieve Vaught’s other testimony if they saw fit.
The other positive evidence that Miller was alive after midnight, December 19th, 1959, was a written statement by Shami Ghandi dated May 5, 1960, which counsel agreed to be introduced in evidence as representing what the witness would have testified if present at the trial. Ghandi stated that he saw Miller after midnight, and that in his opinion Miller left the [564]*564house “Sometime after 1 a.m., Sunday, December 20, 1960.” Plaintiff contends that Ghandi was impeached by the testimony of county investigator Bachus that “Mr. Vaught claimed that he saw them [Moore and Miller] leave around 3 .o’clock, but Ghandi said they were still in the apartment about 3:20,” and further by the impossible statement made by Ghandi that Miller left the house December 20, 1960.
The question arises whether the evidence of Ghandi’s statement could be impeached by his previous statement made to Bachus. Ghandi was not present at the trial. His former statement could not be “called to his mind,” as provided by Code § 38-1803, before it was proved against him.
3 Wigmore Evidence 719, § 1034, states that under such circumstances “it would seem that the rule of prior asking should be dispensed with,” because there has been no opportunity for testing the witness by cross-examination.
The plaintiff did not admit that Ghandi’s statement, introduced by the defendant, was true. The plaintiff clearly could not cross-examine the witness, but it would not be reasonable to construe her agreement to admit the statement in evidence as a waiver of her right to impeach the witness. And defendant in its brief states that Ghandi was open to impeachment if impeaching evidence were available.
The jury could legally have considered Ghandi discredited as a witness if his previous statement to Bachus was in fact inconsistent with his written statement or otherwise impeached.
As a matter of strict logic, the two statements do not contradict each other. But they are inconsistent in the sense that at one time Ghandi stated a definite hour, 3:20 a.m., at which Miller was in the apartment, and later stated with uncertainty that Miller left “sometime after 1:00 a.m.”
To constitute a self-contradiction “It is not a mere difference of statement that suffices; nor yet is an absolute oppositeness essential; it is an inconsistency that is required. . . As a general principle, it is to be understood that this inconsistency is to be determined, not by individual words or phrases alone, but by the whole impression or effect of what has been said or done. . .” 3 Wigmore Evidence 725, '§ 1040.
[565]*565The object of showing self-contradiction as impeaching evidence is to demonstrate the witness’ capacity to make errors, rather than to prove a specific error. “We place his contradictory statements side by side, and, as both cannot be correct, we realize that in at least one of the two he must have spoken erroneously. Thus, we have detected him in one specific error, from which may be inferred a capacity to make other errors. . . It is the repugnancy of the two that is significant.” 3 Wigmore Evidence 685, 686, § 1017.
Because Ghandi’s written statement was introduced and admitted in evidence, we must recognize it alone as Ghandi’s testimony of the truth. The writing prima facie was contradicted by the obvious fact that in December, 1960, Miller was dead. The effect of Ghandi’s statements in and out of court that were in evidence, on Ghandi’s reliability, whether they destroyed his credibility, or were explained as excusable error, we must concede are questions within the province of the jury, and we cannot say as a matter of law that the jury had to believe him. Wilson v. Gray, 34 Ga. App. 320, supra; Lewis v. Patterson, 191 Ga. 348, 357 (12 SE2d 593); Stow v. Hargrove, 203 Ga. 735 (48 SE2d 454). From the verdict it appears that the plaintiff’s evidence, in the jury’s judgment, was of superior weight, enough to incline their minds to plaintiff’s side of the issue as to when Miller died, rather than to defendant’s side. In Goldstein v. Drexler, 102 Ga. App. 90, 91 (2) (115 SE2d 744), it is held: “Where, however, every witness testifying to the fact that there was no consideration for the notes in question was contradicted by other evidence, or by prior contradictory statements on a material point, so that it would be a question for the jury whether any part of the testimony of such witness should be given credit, then the issue should be left to the jury to decide, and their determination as to where the preponderance of evidence lies would be primarily determined by the weight and credit which they saw fit to give to the testimony of the witnesses sought to be impeached.”
In Chapman v. State, 109 Ga. 157, 163 (34 SE 369), the court said: “A witness, for example, may testify to a state of facts the happening of which any person of ordinary intelligence [566]*566would know to be physically impossible. Although no effort might be made to impeach such a witness in any one of the three methods above prescribed, yet no one would contend that a jury would be compelled under the law to- believe a narrative bearing a falsehood upon its face. While the statute prescribes but three methods of impeachment, it clearly does not mean thereby that for no other reason can a witness be discredited. This is too well established by legal principles relating to the credibility of witnesses, and especially the power given by the law to the jury in passing upon the weight to be given human testimony, to require further argument.” Cf. Perry v. Kimberly Jewelry Co., 68 Ga. App. 568, 570 (23 SE2d 471). “A party may show anything which may, in the slightest degree, affect the credit of an opposing witness.” Walraven v. Walraven, 76 Ga. App. 713, 720 (47 SE2d 148).
There was no positive unimpeached evidence in the case that the jury was compelled to believe as against the circumstantial evidence, which was sufficient to support a verdict for the plaintiff. Accordingly, the verdict having been made the judgment of the trial court, this court is not empowered to reverse it. Echols v. Thompson, 211 Ga. 299, supra.
The overruling of defendant’s motion for directed verdict, for judgment notwithstanding the verdict, and for a new trial on the general grounds, was not error.
The defendant contends that a new trial should have been granted because the judge charged the jury that there was no dispute that the amount due, if the defendant were liable, was the amount sued for by the plaintiff. The charge is complained of because the defendant’s answer had denied the allegation of the plaintiff’s petition that the amount payable on the death of the insured was $10,000. The group policy under which the deceased was insured provided for varying amounts of insurance depending upon the nature of the employee’s job and the amount of his pay. There was no evidence as to the nature of Miller’s job nor the amount of his pay. The defendant’s counsel admitted during the closing arguments that, if plaintiff was entitled to¡ recover any amount, she was entitled to recover $15,000.
[567]*567“It is well settled that clients are bound by statements of their attorneys made in open court.” NAACP v. Pye, 96 Ga. App. 685, 686 (101 SE2d 609). Cf. Rogers v. Tillman, 72 Ga. 479; Long v. Lawson, 7 Ga. App. 461 (67 SE 124); Home Finance Co. v. United Motor Sales, 91 Ga. App. 679 (86 SE2d 659). Therefore the charge was not error. Nor was it harmful to the defendants. Ausley v. Cummings, 145 Ga. 750 (89 SE 1071); McKinney v. Woodard, 94 Ga. App. 340, 343 (94 SE2d 620).
In a special ground of its motion for new trial the defendant complains of the admission of testimony by O. J. Kaney, a watch repairman, that the watch found on Miller’s body, and presented to him at the trial, could possibly have stopped running 6 hours and 20 minutes after motion had stopped on a dead man, “if it hadn’t been worn too much the previous day,” or “if he hadn’t been very active.”
This testimony was the witness’ response to a hypothetical question, in which he was asked to assume that the watch was not broken when found on Miller’s body. Defendant contends that this evidence was erroneously admitted because the assumption that the watch was not broken was not supported by facts placed in evidence; and that it was harmful to the defendant because it undermined the circumstantial evidence, that the watch when found was stopped at about 4:20 a.m., supporting defendant’s contention that the accident happened at some time after 3 a.m.
We do not think the evidence objected to was harmful to defendant. Officer Mendenhall testified that he took Miller’s watch at the scene of the accident and carried it around with him until about 2 o’clock in the afternoon (December 20), and it never did start running again. Defendant’s witnesses, officers Mendenhall and Bachus, who testified that the watch was stopped when taken from the body, also testified that so far as they were able to tell the watch was not broken. Neither of them wound it to see if it would start. Mendenhall testified that he did not know whether or not it had just run down. Bachus testified that so far as he knew winding it might have started it again. There was no positive evidence that the watch [568]*568was broken, or that it was not broken, when taken from the body. Officer Bachus delivered the watch to Miller’s father, who' later gave it to Mrs. Miller.
There was no evidence that the watch had been repaired or had not been repaired while in the possession of Miller’s father. There can be no inference one way or the other. Mrs. Miller testified that she had not had it repaired. Her testimony and Mr. Kaney’s was that the watch would run. From all the circumstantial evidence the jury would have been authorized to find that the watch was not broken. The only facts from which an inference could be drawn that the watch was broken are that it was stopped when found and did not start while in Mendenhall’s possession. If any witness had testified as a fact or as an opinion that the watch was broken, defendant’s objection would be less weak.
Since defendant’s own witnesses’ testimony mentioned above, like Mr. Kaney’s answer to the hypothetical question, was that possibly Miller’s watch had just run down, the assumption made by Mr. Kaney, that the watch w'as not broken, apparently did not cause a verdict probably different from that which otherwise would have been rendered, as contended by the defendant. The admission of the evidence, if erroneous at all, was not of sufficient gravity to demand a new trial. Bland v. State, 210 Ga. 100 (78 SE2d 51); McQuire v. State, 29 Ga. App. 192 (114 SE 719); Georgia Ry. &c. Co. v. Shaw, 40 Ga. App. 341 (149 SE 657).
In another special ground of its motion for new trial the defendant contends that the judge’s charge to the jury on the impeachment of witnesses was error “because it instructed the jury that William D. Vaught was impeached if he had merely mistakenly made a contradictory statement before his testimony, whereas the law is that the jury could not impeach William D. Vaught unless the contradictory statement had been wilfully and falsely made by him.” (Emphasis supplied).
The judge did not instruct the jury, either expressly or by implication, that the witness was impeached by a previous contradictory statement mistakenly made by him. Instead, the judge charged: "... A witness may be impeached . . . [569]*569by proof of contradictory statements previously made by him of matters relevant to his testimony and to the case.” This is the language of Code § 38-1803. Comparable charges were approved in Andrews v. State, 196 Ga. 84, 106 (26 SE2d 263); and Weatherby v. State, 213 Ga. 188 (97 SE2d 698).
The charge stated that it was for the jury alone to determine whether or not a witness was impeached by any evidence introduced for the purpose of impeachment.
Therefore, ground 8 of defendant’s motion for new trial is without merit.
Judgment affirmed.
All the Judges concur, except Felton, C. J., Carlisle, P. J., Jordan and Eberhardt, JJ., who dissent.