This is a wrongful death action1 arising out of a railroad crossing collision instituted by the surviving widow and children of O. B. Williams, deceased, against the Fort Worth & Denver Ry. Co. Trial to a jury resulted in a judgment for the plaintiffs, which was affirmed by the Court of Civil Appeals after a portion of the damage recovery was remitted. 367 S.W.2d 925.
The automobile-train collision which claimed the life of O. B. Williams occurred shortly after midnight on September 16, 1956. The deceased accompanied by his minor son was traveling in his automobile on a country highway near the town of Petersburg, Texas. This highway crosses two parallel railroad tracks of petitioner at an angle of seventy-five to eighty degrees. These tracks are several feet apart and run in a north-south direction. The track to the east is the main through track. The west track is a siding. It joins the east, or main, track at two points, one south, and the other north of the highway.
The highway runs west and east across the two tracks. The deceased approached the juncture from the east driving in a westerly direction. The crossing is marked by the standard cross-arm signs. The highway as it approaches the crossing from the east is straight. The terrain surrounding the crossing is level. There are no obstructions to a motorist’s view.
On the night of the accident, a freight train was traveling north from Lubbock to Childress. Behind it and traveling in the same direction was a special passenger train. The freight stopped on the siding to allow the faster passenger train to pass it. As the freight train approached the point where the west side track diverges from the east main track, the brakeman got down and threw the switch allowing the freight to pass onto the siding. After the freight was completely on the siding and stopped thereon, the brakeman threw the switch back to its original position and the passenger train proceeded at a slow rate of speed on the east main track. The fireman on the freight train tossed a fusee, a small flare, onto the highway west of the crossing, since the freight was blocking it. He did not place a fusee oil the highway to the east of the crossing.
When the passenger train was approximately 1300 feet from the crossing, the deceased crossed the east main track and drove into the freight train occupying the ■ west side track. He died a few minutes thereafter. His son knew nothing of the events immediately prior to the collision as he was asleep on the back seat of the automobile at the time.
The engineer and the fireman on the passenger train testified that they saw the deceased’s automobile cross the east main track in front of them and thought that he had safely crossed the west track also as the headlights of the automobile disappeared from view. They said that they assumed that the freight train had been cut *281■or divided so as to leave the crossing clear. Seconds later, when they realized that a collision had occurred, they stopped the passenger train some two to three hundred feet short of the crossing. The passenger train was equipped with the statutory 1500 candle power headlight. It also carried a Mars •oscillating (figure eight) light which was ■operating.
In answer to special issues, the jury found that the crossing was extra hazardous immediately before the collision. Being extra hazardous, it was incumbent upon the railway company to take extra safety measures to protect those using the crossing. The jury found the company has failed in the duty in the following particulars: the freight train crew failed to place a signalman on the highway after the train stopped on the west side track; the freight train crew failed to place a fusee on the highway to the east of the tracks; the freight train crew failed to keep a proper lookout for the deceased and the crew failed to cut its freight train at the crossing. The jury further found that all of these omissions constituted negligence. Accompanying proximate cause issues were found in plaintiffs’ favor. The jury further found that the deceased’s view of the freight train was obscured by the beam of light from the passenger train locomotive shining between the deceased and the freight train, but that it was not negligence for the deceased to continue to drive his car into such beam of light. The jury absolved the deceased of other alleged acts of contributory negligence.
We hold that the trial court erred in receiving in evidence a motion picture film of an experiment or demonstration made by the respondents’ attorney designed to show that a beam of light similar to that cast by a locomotive headlight would cause a “curtain or wall of light” to obstruct the view of a person approaching such beam of light at an angle of approximately ninety degrees.
The questioned experiment was conducted by respondents’ attorney in a public park in the City of Lubbock at about 11:30 at night. To simulate the 1500 candle-power headlight of the passenger train, the attorney placed a police barrel light on top of a pick-up truck. He then positioned his camera to one side of the beam of the barrel light and attempted to photograph objects on the other side of the beam. He moved his camera in a line parallel to the barrel light’s beam, increasing the distance from the barrel light, — the source of the beam. He started exposing the film at about 200 feet from the barrel light and finished the film at a distance of six to seven hundred feet. He also varied the distance of the camera from the beam of the barrel light, from 50 to 250 feet. The objects which he attempted to photograph were located a few feet to the opposite side of the barrel light’s beam. They consisted of two automobiles, a galvanized iron building and some trees.
Counsel for petitioner questioned respondents’ attorney, the sponsor of the motion picture, on voir dire and again on cross-examination as to the conditions under which the experiment was conducted. When asked the power of the barrel light used in the experiment, respondents’ attorney stated that he had been told by the city officials from whom he had borrowed the light that it was 1500 candle-power. Petitioner objected to this testimony as being hearsay and the court sustained the objection and instructed the jury not to consider the witness’ answer. Petitioner then objected to the admission of the film into evidence stating that the experiment had not been conducted under conditions substantially similar to those existing at the time of the collision. This objection was overruled and the jury was allowed to view the film.
In order to render evidence of an experiment made out of court and without the presence of the opposing party admissible, it is generally held that there must be a substantial similarity between condi*282tions existing at the time of the occurrence which gives rise to the litigation and those in existence at the time the experiment is conducted for demonstration purposes. It is not, however, essential that the conditions of the occurrence and the experiment be identical. Houston, East & West Texas Ry. Co. v. Sherman, Tex.Com.App., holdings approved by the Supreme Court, 42 S.W.2d 241 (1931); Rodgers v. State, 93 Tex.Cr.R. 1, 245 S.W. 697 (1922). When there exists a dissimilarity, testimony of an experiment should be excluded when the result thereof would probably confuse rather than aid the jury. There exists an area wherein the determination of the admissibility of proffered experiment testimony rests within the discretion of the trial judge. This occurs whenever the dissimilarity between the occurrence conditions and the circumstances of the experiment is minor or can be made abundantly clear by explanation. Houston, East & West Texas Ry. Co. v. Sherman, supra. Panhandle and Santa Fe Ry. Co. v. Haywood, Tex.Civ.App., 227 S.W. 347, wr. ref. (1921). See, Annotation, Experimental evidence as affected by similarity or dissimilarity of conditions, 85 A.L. R. 479.
The. Court of Civil Appeals was of the opinion that the admissibility of the experiment testimony was a matter within the trial court’s discretion in that the dissimilarities were minor and could be amply explained without confusing or misleading the jury. We are unable to agree with this position. This is not the ordinary photograph case wherein a picture is used to delineate skid marks upon a highway, the physical condition of a road crossing and the like. The purpose of the questioned experiment was to demonstrate the effect of a strong beam of light upon the eyes of a person approaching the beam from an angle of approximately ninety degrees. The motion picture film accordingly represents the retina of the human eye upon the hypothesis that what is shown to be disclosed or obscured on the camera film would likewise be disclosed to or obscured from the human eye. In such experiments, complicated scientific and perhaps physiological factors may be encountered. Care and caution must be exercised in admitting this type of experiment evidence. Dean Wigmore has discussed at some length the dangers inherent in the use of both still photographs and motion pictures of artificial settings. Wigmore on Evidence (3rd Ed.) §§ 798-799, Vol. 3, pp. 201-208. See also, Scott, Photographic Evidence, §§ 624 and 625, Annotations, 27 A.L.R. 913, 19 A.L.R. 877, 83 A.L.R. 1315, 129 A.L.R. 361, 78 A.L.R.2d 152.
Before it can be said that dissimilarity of conditions merely goes to the weight and hence presents a jury question, the judge must be able to say with some degree of certainty, that confusion will not occur and such dissimilarities as are disclosed are capable of explanation so as to be readily understood. In the present case the position of lights, distances and like matters relating to the experiment are not entirely clear. Apparently no exposure of film or camera shot was made across or through the light beam with the origin of the beam (the barrel light) a distance of 1300 feet away, or at a distance approaching 1300 feet. All the testimony we have as to the similarity of the scenes shown upon the film and their actual appearance to the human eye is the statement of an interested nonexpert that conditions disclosed by the film looked about the same to him as the conditions appeared to him when observed at the time the experiment was actually photographed. As above stated, the Diesel locomotive had two lights, one a 1500 candlepower steady light and a Mars oscillating light which may have cast a wider beam than the stationary light and could have lighted up areas on the sides of the track and increased visibility despite the obscuring effect, if any, of the stationary beam. There was a total absence' of a showing of the candlepower of the barrel light used in the experiment. When it appeared that a statement as to the candlepower of the barrel light was based upon hearsay, the trial judge ordered such tcsti-*283mony stricken and instructed the jury not to consider it.
It appears that the petitioner railroad had refused to allow any of its equipment to be used by respondents in conducting an experiment. It also appears that the prime objective sought to be attained by the experiment was to demonstrate that a beam thrown by a light of high candlepower would operate as a “curtain of light” to obscure objects lying behind the beam. The trial judge could properly consider both these factors in ruling upon the admissibility of the testimony, but even so, we are ■of the opinion that the experiment evidence was not sufficient in law to meet the requisites of admissibility. Pittman v. Baladez, 158 Tex. 372, 312 S.W.2d 210 (1958); New York Life Ins. Co. v. Alman, 5th Cir., 22 F.2d 98 (1927); McLendon v. State, 90 Fla. 272, 105 So. 406 (1925). There can be no doubt as to the importance of this experiment testimony to respondents’ case. It was regarded as and no doubt was a highly persuasive piece of evidence. In fact, counsel for respondents in arguing for its admission, despite the dissimilarity of the accident and experiment conditions, asserted that it was his position that “if these (the films of the experiment) are not admissible, then we don’t have any law suit. Now, that’s just the pure common substance of it.” Counsel, in the heat of trial, may have overstated his position, but it can hardly be gainsaid the error of the trial court in receiving in evidence the films of the experiment was reasonably calculated to cause and probably did cause the rendition of an improper judgment in the case. Rule 503, Texas Rules of Civil Procedure, Pittman v. Baladez, supra. The record discloses a reversible error. Casualty Insurance Company of California v. Salinas, 160 Tex. 445, 333 S.W.2d 109, 90 A.L.R.2d 1056 (1960).
It is urged by petitioner as the basis for the rendition of a judgment in its favor that when the improperly admitted experiment evidence is disregarded, there remains no more than a scintilla of evidence supporting the finding that the crossing was extra hazardous at the time of the fatal collision. See, Choate v. San Antonio & Aransas Pass Ry. Co., 90 Tex. 82, 36 S.W. 247 (1896), 91 Tex. 406, 44 S.W. 69 (1898); Joske v. Irvine, 91 Tex. 574, 44 S.W. 1059 (1898); W. St. John Garwood, “The Question of Insufficient Evidence on Appeal,” 30 Tex.Law Review 803; Robert W. Calvert, “ ‘No Evidence’ and ‘Insufficient Evidence’ Points of Error.” 38 Tex.Law Review 359.
It is conceded that the crossing in question is not extra hazardous under normal conditions and that other than providing the warning signs required by statute, the railroad would not be required to take any further measures to protect the traveling public insofar as the crossing involved in this case is concerned. See, Missouri, Kansas & Texas Ry. Co. of Texas v. Magee, 92 Tex. 616, 50 S.W. 1013 (1899); Galveston, Harrisburg & San Antonio Ry. Co. v. Wells, 121 Tex. 310, 50 S.W.2d 247 (1932) ; Texas & New Orleans Ry. Co. v. Hart, 163 Tex. 450, 356 S.W.2d 901 (1962).
Petitioner cites Texas & New Orleans R. Co. v. Compton, 135 Tex. 7, 136 S.W.2d 1113 (1940) and Texas & New Orleans R. Co. v. Stratton, Tex.Civ.App., 74 S.W.2d 741, wr. ref. (1934). These cases hold that the presence of a train on a crossing blocking a public highway docs not in itself constitute an extra hazardous crossing. But the question as to whether a crossing is extra hazardous necessarily depends upon the facts of each case. In Tisdale v. Panhandle & Santa Fe Ry. Co., Tex.Com.App., 228 S.W. 133, 16 A.L.R. 1264, holdings approved by the Supreme Court (1921), it was said:
“Whether or not any given state of facts describing the surroundings of any particular crossing are such as to mark such crossing as one attended with unusual danger or extraordinary hazard is a question solely for the determination of the jury, unless only one conclusion could be drawn therefrom by all reasonable minds.”
*284The Tisdale case is also authority for the proposition that a railroad crossing may present no undue dangers at one time, while at another, it may he extra hazardous. The degree of danger involved depends on the circumstances existing at the time of the accident. With reference to the duty of employing a flagman, it was said in Tis-dale that:
“A flagman might not be required under the law at a certain crossing at one time, and yet it might he negligence to fail to provide one there at another time. The sole question for determination is whether or not at the time of the accident the conditions surrounding the crossing in question rendered it more than ordinarily hazardous or unusually dangerous.”
See also, Missouri-Kansas-Texas R. Co. of Texas v. McLain, 133 Tex. 484, 126 S.W.2d 474 (1939) and Beaumont, S. L. & W. R. Co. v. Richmond, Tex.Civ.App., 78 S.W.2d 232, wr. dis. (1935).
As heretofore indicated, it was respondents’ theory that the beam of light cast by the lights of the passenger train’s Diesel locomotive had the effect of obscuring Williams’ vision and thus rendered the crossing extra hazardous.
While the question is admittedly close, it is the view of a majority of the members of the Court, arrived at after a detailed consideration of the physical circumstances of the collision, the presence of lights and the positions of the trains with reference to the crossing, that when the improperly admitted experiment evidence is disregarded, that which remains raises no more than a scintilla of evidence supporting the theory that the crossing was extra hazardous immediately before the collision occurred. In legal contemplation, a mere scintilla is equated with “no evidence” as it will support nothing more substantial than a surmise or conjecture. See, Joske v. Irvine, supra. However, all members of the Court agree that an order remanding the case rather than rendering judgment should be entered in the interest of justice. Had the experiment evidence been excluded rather than admitted, another species of evidence could probably be resorted to in support of the premise that the lights of the locomotive had an obscuring effect upon the vision of one approaching the crossing. Further, the defect relating to the admission of the evidence above discussed lay primarily with the inadequate basis relied upon to support its introduction. If a proper predicate be laid, photographic evidence could be rendered admissible.
Other matters mentioned in the briefs need not occur on another trial and further discussion is therefore pretermitted.
The judgments of the trial court and Court of Civil Appeals are reversed and the cause remanded to the district court for another trial.
. Articles 4671-4673, Vernon’s Ann.Tex.Stats.