This case arises from the same accident that was considered in Lavallee v. Railroad, 89 N. H. 323, where it was said (at p. 325): “We are unable to discover in the evidence, however, any basis for finding that the failure of the engineer to give the crossing signals was the legal cause of the accident. Under ordinary conditions the driver of the truck would have had ample time and space in which to stop before reaching the track after he discovered the approach of the train [when he was 119 feet from the track]. It is plain that the efficient cause of the collision was the icy condition of the road which prevented the operation of the brakes. It may be argued, however, that if the statutory signals had been given, and had been heard by the driver, he would have stopped before reaching the [second] crest of the grade. The answer is that the burden was upon the plaintiff to prove the causal connection between the negligence complained of and the collision, and the case is bare of evidence upon which to base a conclusion that the conduct of the truck driver would have been different if he had received earlier *326notice of the train’s approach. He made no such claim in his testimony, and the probabilities are all against it.”
In the present cases the driver, thus forewarned, attempted to supply the deficiency in the evidence. Before reciting his new testimony, it will help to its understanding if we describe the contours of the highway upon which the driver approached the crossing. He had first to traverse an up-grade of a length and pitch not disclosed by the evidence. The grade was so great however, and the surface so icy, that he felt obliged to take it at a speed of fifteen miles per hour, “because it was slippery, you know; if I went slower than that, I wouldn’t make it.” It is thus clear that when the truck reached the top of the hill, which may be called the first crest, its speed was fifteen miles per hour.
From the first crest the road sloped gradually for an undetermined distance, then flattened for a very short distance, and finally, from what may be called the second crest, pitched downward more abruptly for a distance, then decreased somewhat, but continued to the crossing. The second crest was 119 feet from the track. It was at this point that the driver first discovered the approaching train and locked his brakes. The truck was then going about twelve or fourteen miles per hour, and entered a slide which carried it to the track, where its front collided with the front of the train. As it reached the crossing the speed of the truck had been decreased to about four or five miles per hour.
The new testimony of the driver was as follows: “Q. Did you know whether or not there was a train due about that time? A. Not at that time; I knew there was a train in the forenoon; if they blew the whistle I planned to stop; I had plenty of chance to stop.” There was no objection to this answer until after the plaintiffs had rested. Having by this unresponsive answer shown his anxiety to make out a case of causation, the witness was questioned briefly about the time, about the up-grade and the succeeding down-grade as far as the crossing. He then testified that he took the up-grade at fifteen miles per hour in order “to make it” and that he knew when he went up what was the succeeding grade to the crossing. Then occurred the following:
“Q. As you went up that hill, what were you doing, in respect to making any effort to find out if a train was coming ? A. If they blew the whistle, I had plenty of chance to stop up at the top of the hill.” Again the anxious unresponsive answer remained without objection until after the plaintiffs rested.
*327“Q. What were you doing to find out if they blew a whistle? A. I was listening. Q. Why were you listening for a whistle? A. Because I knew if a train was coming, I planned to stop---- Q. Why were you listening for a whistle going up the hill? [Objection; no answer]. Q. Why were you listening for a whistle at that point? A. I knew there was a train coming that fore-noon. Q. If the whistle had been blown, and you heard it, when you were going up that grade, Mr. Hebert, what did you intend to do? [Admitted, subject to exception, as relevant to the driver’s state of mind, and then read to the witness with the addition, “What did you have in your mind of doing?”]. A. If I had heard the whistle, I would have stopped.”
Subsequently the driver was asked regarding a previous time when he had been over the crossing. His totally unresponsive answer, “If it blew the whistle that day I had plenty of chance to stop,” was not objected to until after the plaintiffs had rested. The defendant then moved to have the several unresponsive answers stricken from the record, and excepted to the denial of its motions.
The responsive answer, admitted subject to exception, was proper as tending to show the quality of his act. The driver’s state of mind was relevant to the question of his own due care. Furthermore, attention to discover whether a signal was given, plus intention to stop in case it was heard, plus the act of proceeding without stopping, might be relevant to show that no signal was given. Bradley v. Obear, 10 N. H. 477; Wiggin v. Scammon, 27 N. H. 360; Morrow v. Moses, 28 N. H. 95; Moore v. Davis, 49 N. H. 45; Caverno v. Jones, 61 N. H. 623; Smith v. Railroad, 87 N. H. 246, 257. The unresponsive answers of like purport could not be found necessarily to be harmful if the responsive answer was admissible.
But though the matter objected to was superficially relevant and admissible, its admission must be subjected to the primary test of its value. The decision of the Presiding Justice that evidence is admissible depends upon a finding that it is “fit to be considered” by the jury. Its admission, however, is ordinarily subject to the determination of the jury whether it is adequate to establish the fact to prove which it was offered, and the jury may find that “completeness of proof is lacking.” If a finding of incompleteness of proof is conclusively shown upon the record, the testimony should have been excluded or withdrawn from the jury. If the process of preliminary weighing by the court is fully performed, “that which is not worth considering, for one reason or another affecting its value, never *328reaches the auxiliary functionaries, the jurors ... for the Court will of course allow to be considered only such evidence as is worth submitting to men who will judge only by the common and practicable tests .... The judge, in his efforts to prevent the jury from being satisfied by matters of slight value, capable of being exaggerated by prejudice and hasty reasoning, has constantly seen fit to exclude matter which does not rise to a clearly sufficient degree of value, in other words, legal relevancy denotes, first of all, something more than a minimum of probative value. Each single piece of evidence must have a plus value.” 1 Wigmore, Evidence (2d ed.), s. 28.
In the illustrations appended to the passage just cited, the author quotes the following:
“Juries are taken promiscuously from the mass of the people . . . They have generally no previous preparation or possible knowledge of the matter to be tried; and they decide in a space of time too short for any nice or critical disposition. These Judges, therefore, of necessity must forestall the evidence where there is doubt of its competence, and indeed observe much on its credibility, or the most dreadful consequences might follow. The institution of juries, if not thus qualified, could not exist.” Edmund Burke, Report to the House of Commons.
“The very structure, of the system thus produced points to the reason, when we observe its constant, anxious, and over-anxious endeavor to prevent the tribunal to which the evidence is principally addressed from being confused and misled, and from dealing with questions which it has no right to deal with.” Thayer, Preliminary Treatise on Evidence, 2.
Where evidence not “of value” is submitted to the jury, whether because of active error of the Presiding Justice or because the latter has not the time to analyze the content sufficiently, the court of appeal, which is not pressed for time, may analyze it and dispose of it as justice requires. Thus the taking of a voluntary nonsuit has been declared no evidence at all that the plaintiff believed, when he brought the action, that he had no case. Cohn v. Saidel, 71 N. H. 558. When a witness asserts the happening of a physical impossibility, a finding that such a thing occurred will not be sustained. Brown v. Mailhot, 89 N. H. 240. Testimony of a strange and unusual transaction or event, not entitled to credence, will be totally rejected as a matter of law. Moreau v. Insurance Co., 84 N. H. 422. The circumstances under which a false statement was made may be such as to preclude a finding that it was made in good faith. Amoskeag *329&c. Company v. Insurance Co., 88 N. H. 154. In other words, certain evidence which is superficially relevant, may not be the basis for a finding by the jury and is not fit to be submitted to them.
In this case the jury had time and opportunity to note the over-anxiety of George Hebert to testify with regard to a certain material fact. But the jury could have no conception that this anxiety had particular meaning, for they could not know that we had distinctly pointed out in the Lavallee opinion the lack which Hebert was seeking to supply. Assuming that the jury believed Hebert’s denial that he had changed his testimony (the change will later be considered more fully), they believed it in at least partial ignorance of the motive for change which now appears perfectly clear. So they were in a position to be most easily misled.
Still further to mislead them, the jury had not the time, perhaps not the ability, to analyze the record of the testimony at the former trial and by it test the credibility of his new testimony. Probably the Presiding Justice had not the time to analyze the complicated factors presented. We have therefore done so in order that it may be determined whether any misleading, incredible or valueless testimony was given to the jury in such manner as to affect the verdict.
At the outset, it is to be remarked that if George Hebert had on the day of the accident a state of mind such as he now asserts, it was so essential to make a case for Lavallee, for himself and for his brother, that it is difficult, if not impossible, to assign to inadvertence his failure to give testimony in the Lavallee case which he has now volunteered repeatedly when not required by the questions addressed to him. The real key to this is the motive produced by our former decision, known to him and not to the jury.
There was another important change in his testimony, which the jury had insufficient opportunity to analyze. At the Lavallee trial, George Hebert testified that after he locked his brakes, he attempted to steer into the snow-bank at his left; that he would have succeeded but for some ruts in the ice; that his rear wheels skidded out of the ruts to the right, but his front wheels remained in the ruts and defeated his purpose. When he so testified, certain photographs had been introduced showing the locus of the accident when there was no snow or ice on the ground, and there was as yet nothing to throw doubt upon his story. After his testifying thus, the plaintiff produced two small photographs taken by an amateur immediately after the accident. Their size is such that they may not be of much impor*330tance; but at least they show no ruts. Still later the defendant introduced (of course after George Hebert had testified), three large photographs taken the day after the accident. These showed all of the road for a distance of nearly 300 feet from the crossing. There is no vestige of a rut, but an entirely smooth road, except for a possible trace of a rut within a few feet of the crossing, so slight as not to have any probable effect.
The driver laid no emphasis upon ruts in the trial of his own case. Being then faced with the defendant’s photographs, he admitted that they correctly represented the conditions on the day of the accident. He further admitted that no ruts were discernable in a photograph taken 93 feet from the nearest rail.
But circumstances impelled the driver to change his testimony in another respect. At the first trial, he testified that the right-hand window of the cab was closed; the left-hand window open one-half an inch. At the second trial, he said that the latter window was opened an inch. There is small difference in a half-inch, but an opening multiplied by two would obviously be capable of admitting sound more readily, and would be more consistent with the definite intention which was brought forward for the first time at the last trial.
The driver also categorically and in detail denied every statement he had given to the defendant’s claim agent—-unnecessarily so, because nothing in the statement contradicted him in any material respect. He denied that he had made any statement at all. There can be no doubt that he did make one. The other plaintiff first, had a lapse of memory as to his statement, then admitted that the claim agent questioned the plaintiffs and that they answered, and ended by denying that they were questioned at all.
Upon the question of causation, everything depends upon the driver’s testimony. We regretfully say that in our view it is unreliable. As to the possibility of hearing the whistle in time to avoid the accident, the relative positions of train and truck were of the first importance. At the first trial the driver said that when he first saw the train it was “just this side of the big pine tree,” “down near the pine tree” located some more than 700 feet from the crossing. The train could have been seen by the driver when 600 to 700 feet distant from the crossing. By the law of the trial the whistle should have been sounded at 825 feet. Assuming that the distance was 600 feet and allowing five seconds for the train to go 225 feet, the truck would have been only 219 feet from the crossing, on a *331down-grade, when the time for signalling came. That would seriously trouble the plaintiffs, who wished the signal to sound when they were on the up-grade, over 350 feet from the crossing, so that they could clearly stop before the truck was going down-hill on the ice.
That necessitated locating the train, when first seen, considerably nearer to the crossing than the pine tree by which the witness had verified the observations testified to at the first trial. In order to place his truck on the far side of the hill when the signal was sounded, so that a listening ear would not hear it, he abandoned testimony checked by a fixed object at a known distance, and asked credence for new testimony absolutely without corroboration.
At the last trial George Hebert testified on direct examination thus: “Q. When you saw the train where was it? A. When I took the down grade. Q. No, where was the train then? A. It was about just this side of the pine tree, around 400 or 500 feet.”
Upon cross-examination at the last trial, the following occurred: “Q. Hebert, when you first saw the train, the train was about opposite the pine tree down there, shown on the plan, wasn’t it? A. This side of the pine tree, (indicating on plan) Q. About opposite? A. I said this side of the pine tree (indicating), around there (indicating). Q. How much? A. Four hundred or five hundred feet. Q. Four or five hundred feet this side of the pine tree (indicating)? A. Yes, sir. Q. The pine tree is 700 feet from the crossing, and if it was 450 feet this side of the pine tree—A. I said— Q. (Concluding question) It would be 250 feet from the crossing? A. Oh, no; I made a mistake on that; I said 400 to 500 feet to the crossing, this side of the pine tree (indicating). Q. That is what you mean? A. Yes, sir. Q. You testified in this case, in the Lavallee case, in November, 1936? A. Lavallee was all alone in his case; I was just a witness. Q. You were a witness in that case and testified to the facts in this case, in this happening, you testified in November, 1936, about this accident, in the Lavallee case? A. I was just a witness. Q. Are you any better than a witness today? A. Yes; I am fighting the case myself, today. Q. So it makes a difference in your testimony, does it? Yes or no. A. No, sir. Q. Well, why didn’t you say, in November, 1936, about where the train was when you first saw it? A. I told the same thing; 400 or 500 feet from the crossing; 400 to 500 feet from the crossing.”
The witness was not badgered. When he was asked, “you testified in this case, in the Lavallee case, in November, 1936,” his quick answer shows beyond any doubt consciousness that he had changed *332his testimony to meet the opinion in the Lavallee case. When he later declared under oath that he “told the same thing; 400 or 500 feet from the crossing,” he must have known that he was telling an untruth. He may be taken at his own word, “I am fighting the case myself, today.”
The jury in this case were given the record of the evidence of George Hebert at the former trial, “to prove that Mr. Hebert did not testify on certain points at this other trial.” This we understand was for the general purpose of contradiction and for the bearing it had upon the credibility of the new testimony.. While the credibility of such testimony is ordinarily for the jury, we think that no reasonable man could credit this.
It is true that with very few exceptions a trier of facts may accept some parts and reject other parts of the testimony of a witness, and may adopt one or the other of inconsistent statements of the witness. But credibility may become a matter of law. “Testimony of what may be possible may fall short of being entitled to credence as probable. There are limits to a fair and honest belief of testimony of a strange and unusual transaction or event, and the fact of its possible truth, may not be enough to give it acceptance to the ordinary man. When in addition the testimony is shown to be inconsistent with that of the deposition in important points, the difficulty of reasonable acceptance of the testimony becomes greater.” Moreau v. Insurance Co., 84 N. H. 422, 425. Here, the interest of the driver as a party so glaringly colored his testimony, that at best no fair minded man could balance it and find it true. If not necessarily false, it did not tend to show the truth. The combination of interest, contradiction and preparation is so strikingly displayed that a reasonable conclusion of the truth of the testimony may not be drawn when no supporting evidence upon an essential issue is introduced. Justice does not countenance it.
“It is not to be doubted that he appreciated the purport and bearing of the questions asked him. They were not ambiguous and with or without reference to the context could not be misunderstood. To say that he might reasonably be found honest in his testimony would exhibit an indifference to the record that would permit a clear case of dishonesty to be toned down into honest mistake. On a fair reading of the record no such result of an honest mistake can be reached. While allowance for misunderstanding may be given, irresponsibility for testimony may not be granted.” Moreau v. Insurance Co., supra, 426.
*333The uncorroborated new testimony of George Hebert, when placed beside his inconsistent corroborated testimony at the earlier trial, furnishes no adequate basis for the finding of facts essential to make a case for the plaintiffs. On the contrary, it was so obviously “capable of being exaggerated by prejudice and hasty reasoning,” that the jury should not have been permitted to consider it. Without the new testimony the present case stands upon the same footing as the Lavallee case. There is no reason to believe that the driver’s conduct would have been different if he had heard the whistle a few seconds earlier than he did. He saw the train at the point where it first became visible from the highway. He immediately applied his brakes, and the accident occurred because of the icy condition of the highway. There is no basis in the evidence for a finding that he would have applied them any sooner if the whistle had been blown, or, if he had applied them sooner, how much sooner and with what effect upon the progress of the truck over the icy roadway.
Judgments for the defendant.
Branch and Woodbury, JJ., dissented: the others concurred.