This is an action for wrongful death. Plaintiff had a verdict and judgment in the amount of $40,000 and defendant appeals.

Plaintiff’s intestate, Steven Eader, was killed on June 11, 1969, at about 8:30 p.m. while he was driving his car in an easterly direction on Highway 212 in Clackamas county in the vicinity of West Linn. A rock entered the windshield of his car and apparently struck him on the head, causing death. There were no eyewitnesses who saw the rock strike the car. A new highway was under construction in the area at the time, and defendant was the general contractor for the project. In connection with its construction activities, defendant had constructed a “haul road” near Highway 212. The haul road ran roughly parallel to the highway in the area of the accident, and lay above the highway on a hillside. It was plaintiff’s theory that the rock which killed Steven Bader rolled or fell from the haul road. Defendant contends that plaintiff failed to prove that theory, and also relies on evidence to support its own theory that the rock was flung by the tires of a truck which passed Eader’s car on the highway.

*358At the close of all the evidence, defendant moved for a directed verdict on the ground that the evidence did not show that the rock fell from the haul road. The motion was denied. Defendant also moved the court to strike from the complaint all the specifications of negligence. This motion was granted as to an allegation that defendant was negligent in using a haul road in dangerously close proximity to the highway, but was denied as to the other two specifications of negligence — failing to maintain a barricade or other protective device to prevent rocks from falling onto the highway from the haul road, and failing to provide traffic controls on the highway. Defendant also excepted to the trial court’s instructions submitting the two specifications of negligence to the jury. The first four assignments of error challenge these rulings and raise the question whether there was sufficient evidence of causation and of negligence.

We find no error in these rulings. There was ample evidence from which the jury could find that the rock came from the haul road above the highway, and that defendant could have prevented the accident had it maintained an adequate barricade or employed some form of traffic control to stop highway traffic while equipment was moving on the haul road. There was testimony that, at the time of the accident, large trucks or construction vehicles were using the haul road. A policeman who regularly patrolled the highway testified that large trucks moved continuously on the haul road and that he had, on occasion, clocked the trucks at speeds up to 40 miles per hour while they were loaded with rock.

There was also evidence which tended to show that rocks had fallen onto the highway from the haul *359road on earlier occasions and that defendant was aware of this.

The haul road had a rock and gravel surface. Along the side closest to the highway there was a “berm” or ridge of loose rocks of various sizes. One witness testified that the berm averaged approximately two feet in height; pictures of the haul road show the berm, which appears somewhat lower than two feet in many areas. No other barriers or protective devices were provided, and no flagmen or traffic signals were employed at the time of the accident.

Other assignments of error challenge the admission of testimony that rocks had been seen on or falling to the highway on prior occasions. Evidence of prior similar occurrences is admissible under some circumstances in a negligence action. As a general rule, evidence of prior accidents or acts of negligence is not admissible to prove a specific act of negligence. Saunders v. Williams & Co., 155 Or 1, 7, 62 P2d 260 (1936). Such evidence is, however, admissible to prove the existence of a continuing defect or a continuing course of negligent conduct, and that the condition or course of conduct is in fact dangerous, or that the defendant had notice of its dangerous character. Clary v. Polk County, 231 Or 148, 152, 372 P2d 524 (1962); Saunders v. Williams & Co., supra, 155 Or at 7; Krause v. Southern Pacific Co., 135 Or 310, 316, 295 P 966 (1931); Sheard v. Oregon Elec. Ry. Co., 131 Or 415, 425, 282 P 542 (1929); Coates v. Marion County. 96 Or 334, 340, 189 P 903 (1920); Gynther v. Brown & McCabe, 67 Or 310, 319, 134 P 1186 (1913). The admissibility of such evidence for these purposes is, however, subject to the requirement that the prior accidents must have occurred under similar conditions and circumstances. Galvin v. Brown & McCabe, 53 Or 598, 614, 101 P 671 *360(1909); Lee v. Meier & Frank Co., 166 Or 600, 610-611, 114 P2d 136 (1941). In the Lee case the court said that ascertainment of similarity of conditions may ordinarily be left to the discretion of the trial judge. The general rules are discussed and the cases collected in an Annotation, 70 ALR2d 167.

In the present case the dangerous character of a continuing condition was clearly at issue. Plaintiff contended that defendant was negligent in failing to take precautions to protect against the danger of rocks falling to the highway from above. The evidence objected to tended to show both the existence of a dangerous condition and that defendant had notice of the dangers. The real question is whether the circumstances were sufficiently similar in each instance.

Only substantial similarity, not complete identity of circumstances, is required. See 70 ALR2d at 201 et seq. What elements must be similar will depend, of course, on the nature of the allegedly dangerous condition in each case.

In this case the Rader accident occurred at about 8:30 p.m. on June 11, 1969, during daylight hours. There was evidence that construction vehicles were on the haul road at the very time of the accident. The berm was in place.

The testimony of Officer Irwin was that during that month of June he traveled the section of the highway frequently, and that many times he saw rocks on the highway while defendant was working. He also testified, however, that he never actually saw any rocks fall except during blasting operations, and that flagmen were employed during blasting. He did not pinpoint any specific spot on the highway where he had seen rocks.

*361Comer testified that his windshield was struck by a rock, the source of which he did not know, in the morning hours of a day early in May. There was equipment working on the haul road at the tune. The jury could have found that this happened very near the scene of the Rader accident, and that the berm was in place at the time.

The time, place, and circumstances of these incidents are sufficiently similar to those of the Rader accident, and admission of this evidence was not error. The evidence tends to prove that the movement of equipment on the haul road created a danger of accidents like that suffered by Rader. Since Mr. Comer also testified that he notified defendant about the rock which hit his windshield his evidence is directly relevant on the question of defendant’s notice of the dangerous condition.

Witnesses Peter and Knapp testified that in April they saw a boulder roll down across the highway from the area of the haul road. Mrs. Knapp testified that the matter came to defendant’s attention. The haul road was under construction at the time, and the boulder was apparently dislodged by a caterpillar engaged in the road construction. The boulder crossed the highway about a quarter of a mile from the scene of the Rader accident. Mrs. Knapp also testified that she saw other rocks on the highway during the month or two prior to Rader’s death, at times when construction equipment was working on the haul road.

Although some circumstances of the boulder incident are different from those surrounding the Rader accident, we think the evidence was properly admitted. The physical relationship between the haul road and the highway, a very significant factor, remained the same. The evidence tended to show that rocks which *362were dislodged by activity in the area of the haul road were a danger to highway traffic, and that defendant was aware of this danger.

Defendant suggests that the admission of evidence of prior occurrences is rendered erroneous by the absence of an instruction limiting the use of the evidence to its proper purpose. Such instructions were mentioned and approved in Clary v. Polk County, 231 Or 148, 152, 372 P2d 524 (1962) and in Coates v. Marion County, 96 Or 334, 189 P 903 (1920). It was defendant’s duty, however, to ask for such an instruction, and the trial judge is not required to give such an instruction unless requested. Eleanor Walls v. Clark, 252 Or 414, 417-418, 449 P2d 141 (1969). Defendant made no request for a limiting instruction.

Two assignments of error concern the trial court’s refusal to permit defendant’s expert witness, Ronald McReary, to testify to certain opinions. Mc-Reary is a civil engineer with a master’s degree in engineering. Since 1964 he has headed a consulting engineering firm which operates in Oregon, Washington, Idaho, and California. His prior experience includes two years as a member of the Oregon State University engineering faculty, and more than ten years spent in “the administration and supervision of heavy construction work, doing safety analysis, project management on heavy construction projects.” He was employed by defendant to investigate the Rader accident; he examined the accident site, the Rader vehicle, and the rock which struck it.

Defendant’s offer of proof, which the trial court rejected, was made in the form of a summary by counsel. McReary would testify, he said, that in his opinion the rock which entered Rader’s ear had been *363between dual tires. This opinion was based on the fact that the rock was worn on one surface, and that there were traces of tire rubber on two sides of the rock.①

Counsel stated that McReary would also testify in his opinion the rock had flown from between dual tires and entered the Rader automobile. This opinion was based on the fact that the diameter of the rock was the same as the distance between the dual tires of a truck which, according to the truck driver, had passed Rader’s car at about the time of the accident, and, on McReary’s opinion, already in evidence, that when the rock entered the car it was traveling on a rising trajectory.

As to the first opinion, we think no proper foundation had been laid. Although McReary testified that from his experience on construction projects he knew that rocks were occasionally picked up between dual tires, he did not testify that he had ever examined such rocks or that he had any special competence in determining from a rock’s appearance whether it had, in fact, been between dual tires.

However, even if we assume that the witness was properly qualified to express an expert opinion on that subject, the offer of proof went further, also including McReary’s opinion that the rock had been flung into the Rader car from between dual tires.②

*364In addition to the facts stated in the offer of proof as a "basis for this opinion, other relevant evidence was before the jury. Robert Strike, a truck driver for one of defendant’s subcontractors on the highway project, testified that on the day of the accident he was driving a track with daal wheels on the highway. He had jast picked ap a load of crashed rock, had come down the haul road onto the highway, and had headed west. Rader’s ear, going east, passed his track and, when the car was opposite the cab of the track, Strike heard a noise which he described as a “load crash”. When he returned he saw the Rader ear where it had come to rest, and he testified that that point was approximately 200 to 300 feet from the point where he had passed the car and heard the “load, crash”. However, Michael Hedge, who was driving along the highway in the same direction as Rader and who saw the Rader car both shortly before and .shortly after the impact with the rock testified that he had no recollection of any westbound vehicles at that time. He could not testify for certain that there had not been any westbound traffic, however.

The offer of proof did not state that the witness’s opinion was based on the assumption that Strike’s track passed the Rader car at the crucial time, and that Strike heard a load noise as he passed the car. However, the witness’s assumption that Strike’s track passed the Rader car is evidenced by his reliance on the fact that the width of the rock was the same as the distance between Strike’s track tires. As we have seen, the testimony of the witness Hedge, while not directly contradictory, did east some doubt on Strike’s testimony. The jury was entitled to disbelieve it.

Although the hypothetical state of facts upon *365which an expert opinion is based may include any combination of assumed facts which are supported by substantial evidence, such assumptions, especially where the evidence is conflicting, should be clearly stated. The witness should not be allowed to express an opinion based on an unarticulated assumption that certain controverted evidence is true. Carruthers v. Phillips, 169 Or 636, 645, 131 P2d 193 (1942); Lippold v. Kidd, 126 Or 160, 168, 269 P 210 (1928); Lehman v. Knott, 100 Or 59, 70, 196 P 476 (1921). There are also statements in these cases that a witness may not invade or usurp the province of the jury. We have rejected the notion that evidence may be objectionable as an invasion of the jury’s function. Ritter v. Beals, 225 Or 504, 524-525, 358 P2d 1080 (1961). Thus opinion testimony may be admissible even though it concerns an ultimate fact to be found by the jury. However, we consider the cited cases to be still sound insofar as they hold that a question calling for an expert opinion must disclose the critical assumptions upon which that opinion is based. As the jury must pass upon conflicting evidence, and also must evaluate expert opinion testimony in light of the validity of the facts and assumptions upon which it is based, the expert’s assumption of the truth of a controverted fact should be made clear to the jury. The trial court did not err in rejecting the offer of proof.③

*366The final assignment of error concerns a matter not raised in the trial court in any form, and we decline to notice it.

Affirmed.

concurring.

The disagreement between the majority and dissenting opinions is whether the trial court was guilty of reversible error when it would not allow an engineer to testify that, in his opinion, the rock which killed decedent had been between dual tires and that it had flown from between such tires when it entered decedent’s vehicle. The following testimony (virtually all of which came from the expert) was allowed by the trial court to go to the jury:

1) the width of the rock;

2) the width of the space between the dual tires of the truck, which some evidence indicated passed decedent’s vehicle at the time of the accident;

3) that the space between the tires and the width of the rock were the same;

4) that the dark marks on two sides of the rock were rubber marks;

5) that the rock was flat from wear on a third side;

6) that the chrome at the top of the windshield and the rod supporting the sun visor were bent upward above the level of the top of the windshield ;

7) that the rock was on a rising trajectory when it entered the vehicle “as if it had started at the ground and was rising upward”;

8) that dual tires pick up rocks and throw them and that this is why some trucks have “rock bars” installed between their dual tires.

*367Under such circumstances, I am not very concerned about the fine legal problem of whether the expert should be allowed to testify that the rock was thrown through the windshield by the dual wheels of a truck. The witness was allowed to show the jury in detail exactly how the defendant claimed the accident occurred and to use his expertise in explaining why it was so contended. The important part of his testimony was that by which he used his expertise to show the jury how the rock had to have been on a rising-trajectory for the damage to the vehicle to have occurred as it did. In view of all this, it is my opinion that there is insufficient likelihood that the trial court’s refusal to allow the witness to express his ultimate opinion had any effect on the outcome of the case, and, therefore, a reversal is not justified.

dissenting.

The majority opinion justifies the rejection of McReary’s testimony on the ground that he was not qualified as an expert witness since he did not testify that he had ever examined rocks which had been picked up between dual tires or that he had any special competence in determining from a rock’s appearance whether it had in fact been between dual tires. However, plaintiff’s counsel did not object to the evidence, nor did the trial court exclude it, on that ground. The trial court gave no indication whether it regarded McReary as sufficiently expert to testify that in his opinion the rock in question had been between dual tires; however, the record contains substantial evidence as to McReary’s general expertise on highway safety and engineering, and McReary testified that he knew from professional experience of rocks being *368picked up between dual tires. There may be cases in which a question is so obviously beyond the scope of a witness’ expertise that we would hold that the trial court would abuse its discretion if the witness were permitted to testify, even absent an objection from the opposing party. The majority apparently believe this to be such a case. I disagree.

A witness is not disqualified as an expert simply because he has not previously observed the specific cause and effect relationship sought to be shown in the case in which he is a witness. He is entitled to draw upon the broader principles which form the foundation for his expertise and reason from those principles to explain the specific cause and effect connection in the specific case. That was what McReary purported to do in the present case.

As an alternative ground for excluding the evidence, the majority opinion holds that the form of the hypothetical question was improper in that it failed to state that the witness’ opinion was based on the assumption that Strike’s truck passed the Rader car at the crucial time and that Strike heard a loud noise as he passed the car. I am unable to understand why that assumption is a necessary part of the hypothetical question. It was McReary’s conclusion that the rock was on a rising trajectory as it entered the car, having been thrown from between dual tires of a passing truck. This conclusion could be reached reasonably without a consideration of Strike’s testimony relating to the noise he heard as he passed the Rader car.

But assuming the hypothetical was defective in form, it was not excluded by the trial court for that reason. The trial court said:

“THE COURT: Now, I am not going to let this *369witness testify an opinion as to how this rock could have gotten thrown at the windshield of this car. * * * This is the same proposition as having someone go out and investigate an automobile accident and tell us how it happened. They haven’t permitted that in this state yet.”

Thus it is apparent that the evidence was excluded on a ground that had nothing to do with the form of the question. It appears that the trial court felt that the witness was invading the province of the jury in drawing his conclusion as to the manner in which the accident occurred. Although the jury did not have to accept this opinion, it would have been helpful if accepted. Consequently, I think that the trial court’s reason for excluding it was erroneous. Therefore I would reverse and remand for a new trial.

In his concurring opinion Mr. Justice Holman takes the position that even if the trial court erred in not permitting McReary to testify, the court’s action was not sufficiently prejudicial to constitute reversible error. I disagree.

It must be borne in mind that the rock which penetrated the windshield of decedent’s car was very large and heavy, with dimensions of approximately nine inches in length and four and one-half to five inches in width. Unless someone with scientific knowledge explained to a juror that it was possible for a rock of that size to be picked up and thrown through a windshield, it is likely that the jury would not believe that it could happen. It is to be noted that the testimony explaining that “the rock was on a rising trajectory when it entered the vehicle,” left unexplained, was equally consistent with the theory that the rock came from the haul road above and bounced from the pavement into the windshield. The expert *370witness should have been permitted to explain that in his opinion the rock was thrown by the truck tires and did not rise from the ground in some other way.

I question the statement in footnote three of the majority opinion, that “where the ruling of the trial judge is correct on any ground it will be upheld, even though his reasons for the ruling may have been erroneous.” That rule is applicable where a judgment is rendered on an incorrect ground but there is a correct ground which will sustain it. However, the principle is different where the court rules on the admissibility of evidence. In the latter case if a specific objection is made and sustained on an improper ground, the exclusion will not be upheld on appeal if it appears that had the proper ground been given the proponent of the evidence could have obviated the objection.①