Challis Irrigation Co. v. State

689 P.2d 230 (1984) 107 Idaho 338

CHALLIS IRRIGATION COMPANY, an Idaho corporation, Glendon Hunt, Gary Chamberlain, and Ralph Anderson, Plaintiffs-Respondents,
v.
STATE of Idaho, Defendant-Appellant.

No. 13531.

Court of Appeals of Idaho.

May 22, 1984. Rehearing Granted August 10, 1984. Upon Rehearing October 15, 1984. Petition for Review Denied December 21, 1984.

Jim Jones, Atty. Gen. by Stephen V. Goddard, Deputy Atty. Gen., and William B. Dillon III, Idaho Dept. of Fish and Game, Boise, and W. Marcus W. Nye, Racine, Olson, Nye, Cooper & Budge, Pocatello, for appellant; Stephen V. Goddard argued the appeal; William B. Dillon filed a brief in response to the petition for rehearing.

Dwight E. Baker, St. Clair, Hiller, Wood & McGrath, Idaho Falls, for plaintiffs-respondents.

ON PETITION FOR REHEARING

This opinion supersedes our prior opinion issued January 10, 1984, which is hereby withdrawn.

BURNETT, Judge.

This case presents a question of causation. We are asked to decide whether there is substantial evidence to support a jury finding that improper maintenance of a fish screen in an irrigation canal caused a break in the canal wall. This issue is raised by the State of Idaho, Department of Fish and Game, which installed and maintained the screen under a license granted by the Challis Irrigation Company, owner of the canal. The State had a contractual duty, under the terms of the license, to avoid obstructing the flow of water through the fish screen. The canal company and several of its member farmers sued the State, alleging that an obstruction at the fish screen had caused water to back up, breaking the canal bank. Upon verdicts in favor of the canal company and farmers, the State was adjudged liable for the resulting damage.

This case has proven vexatious on appeal. The causation issue, though simple in concept, is complex when viewed against an intricate — and in some respects enigmatic — record. In a prior, unreported opinion, we initially held that the jury's verdicts were unsupported by substantial evidence of causation. Judgments upon the verdicts were reversed. In response to a petition for rehearing, we have reexamined the entire record and have studied the detailed, supplemental briefs submitted by both sides. The causation issue has been presented exhaustively, and further argument upon rehearing is unnecessary. Upon reconsidering our prior opinion, we now conclude, for reasons set forth below, that the verdicts should be upheld.

In Part I of today's opinion we broadly outline the rules governing appellate review of the trial record in an appeal from judgments upon jury verdicts. We then set forth the pertinent facts from the record relating to the issue of causation. In Part II we focus upon the legal tests of proximate cause and upon the use of circumstantial evidence to meet those tests in this case.

I

Appellate review of a jury verdict is constrained by several well settled principles. The general principle, initially articulated by I.C. § 13-219 (now repealed) and perpetuated by case law, is that a jury verdict will not be overturned if it is supported by substantial, although conflicting, evidence. E.g., Quincy v. Joint School Dist. No. 41, 102 Idaho 764, 640 P.2d 304 (1981). From this principle flow a number of closely related corollaries. A question of proximate cause is for the triers of fact to determine, and their determination will not be disturbed on appeal if supported by substantial evidence. Mann v. Gonzales, 100 Idaho 769, 605 P.2d 947 (1980). Where there is conflict as to the meaning of, or reasonable inferences to be drawn from, the evidence, the triers of fact should make the determination. Pierson v. Jones, 102 Idaho 82, 625 P.2d 1085 (1981). On appeal, the evidence in the record, and all reasonable inferences, are viewed in favor of the jury verdict. E.g., Henderson v. Cominco American, Inc., 95 Idaho 690, 518 P.2d 873 (1974).

With these constraints in mind, we turn to the record in the present case, noting the evidence which is undisputed or — if disputed — reasonably could have been viewed in favor of the irrigation company. The company's canal starts at a headgate on the Salmon River and runs at a slightly declining grade into a farming area. The fish screen is located in the canal roughly two miles below the headgate. The structure includes a trash rack in front of the screen, to catch debris floating down the canal. Approximately fifty feet in front of the trash rack, and along one side of the canal, the State has constructed an opening into a bypass channel which loops around the trash rack and fish screen, rejoining the canal on the downstream side. This bypass opening contains float-activated gates designed to open if clogging in the trash rack impedes the flow in the canal, causing the water to rise.

The canal break in question occurred during the night of June 15 or in the early morning hours of June 16, 1974. During June the Salmon River rose to a record flood level. One of the canal company's witnesses testified that the river "had been on an increase for approximately five days" before June 16 and had "pretty well stabled out" on the 16th. When the river ran high, the canal headgate became filled with debris; but the river poured over or around the headgate and entered the canal. Although the precise quantity of water flowing in the canal before the break was not measured, it was estimated to be 175 cubic feet per second (cfs) or more — perhaps substantially more. A witness for the canal company later testified that the flow was greater than normal. The canal company could have used "spillboards" below the headgate to divert some of the water back to the river, but it did not do so.

The increased flow in the canal was accompanied by floating debris from the river. The debris began to accumulate against the trash rack in front of the fish screen. On June 14, a work crew from the State was dispatched to the scene. Four men worked all morning, and two men worked that afternoon, to clear the debris. On June 15, a crew of two men arrived. They noticed that debris had accumulated back to the bypass gates and was coming in about as fast as they could remove it. They also noticed that the water level in front of the trash rack was a few inches higher than the level below the fish screen. After clearing debris from about half the width of the channel, they quit and went to work elsewhere for the rest of the day.

That same afternoon a farmer, whose property was located down the canal from the fish screen, measured the water flow against a peg device on his property. He testified at trial that he found the flow to be normal. He also testified that late at night, he rechecked the peg, finding the water level to be somewhat below normal. The following morning, June 16, farmers in the area noticed that their irrigation systems were not working. Upon investigation they found that the canal had broken upstream, nearly a mile above the fish screen. A witness later testified that the break occurred at a "low point" in the canal — that is, at a place where the canal had the least distance between a normal water level and the top of its banks.

One observer who viewed the canal above the break point noted that "quite a lot of water" was flowing in the canal between the river and the break point on June 16; but he added, "I have seen more water than that in that canal." On the downstream side of the break point, farmers discovered a large amount of debris spread upon the bottom of the empty canal, directly in front of the trash rack. The debris occupied the full width of the canal and extended back to the bypass gates. An observer later testified that he saw silt accumulation along the canal banks just upstream from the trash rack twenty-one inches, or about 1.7 feet, higher than the normal water level. Because the normal depth was approximately 3.6 feet, this testimony suggested that the water in the canal had reached a depth of about 5.3 feet near the trash rack before the break occurred upstream. The same observer said he also saw a silt line near the break point "a good fifteen inches" above the normal water level.

At the bypass, one of the float-activated gates was found partially open. The others were found shut. The record does not indicate the exact water levels at which the floats were designed to open these gates; but the gates themselves were only 4.3 feet high. Presumably they were designed to open at a somewhat lower water level. An obvious anomaly, never explained at trial, was how the water could have reached a depth of 5.3 feet near the trash rack without tripping all of the bypass gates. There was no evidence that the gates had malfunctioned. On the other hand, although the gates had been visually checked, they had not been tested on or immediately before June 15. One of the State's witnesses explained that the gates were not routinely tested because any such test would cause a loss of fish into the bypass channel.

Upon this circumstantial evidence, the canal company and the State asked the jury to draw competing inferences. The canal company asserted that during the night of June 15-16, the trash rack had become so clogged that the water in the canal backed up; and that the increased depth attributable to clogging, when added to the existing flow, caused a canal bank to fail at the "low point" where the washout later was discovered. In contrast, the State contended that debris at the trash rack added little, if at all, to the water depth at the break point. Rather, the State urged that an abnormally high quantity of water entering the canal from the river caused the water level to rise until the canal bank failed.

To aid the jury in evaluating the evidence, the parties presented opinion testimony by two expert witnesses. The canal company's expert exhibited a diagram showing that if the water near the trash rack had increased to the level of 5.3 feet, and if the water flow at that time had been approximately 175 cfs, the water upstream would have reached a depth approximately equal to the height of the canal bank at the point of the break. The expert did not purport to know, in fact, why or whether the water had reached 5.3 feet near the trash rack; he simply assumed that it had. He acknowledged that if this depth was attained, the primary reason likely would have been the quantity of water entering the canal from the river. But he stated that if the trash rack had been clogged, such clogging could have contributed approximately two to three inches to the depth of the water at the break point upstream. The State's expert opined that virtually all of the increased depth at the break point would have been due to the high intake of water from the river, and no significant amount would have been caused by the alleged clogging downstream.

Upon this evidence, the jury returned verdicts for the canal company and farmers against the State of Idaho. The jury implicitly inferred that the trash rack had become clogged, obstructing the flow of water in the canal. This is a reasonable inference, supported by substantial evidence. The jury further inferred, implicitly, that the clogging caused the canal bank to fail. The latter inference is the point of controversy in this appeal.

The State argues that the trial court should have granted a directed verdict or post-verdict relief, and that in any event the verdicts should be set aside on appeal. The State's arguments embrace a common question — whether there is substantial evidence to support the inference that the State's failure to prevent obstruction to water flow at the trash rack caused the damage to the canal. We now focus upon that question.

II

As noted above, the State's failure to prevent an obstruction represented a breach of a contractual obligation imposed by the license for the fish screen. In order to recover damages for a breach of contract, the aggrieved party must show that his loss actually resulted from the breach. E.g., Graham v. Asbury, 112 Ariz. 184, 540 P.2d 656 (Ariz. 1975). In this case the trial court chose to frame the causation issue in terms of "proximate cause." The concept of proximate cause contains two components — cause in fact, and scope of legal responsibility. W. PROSSER, HANDBOOK OF THE LAW OF TORTS §§ 41-42 (4th ed. 1971). Here, our focus is upon cause in fact.

Proximate cause, in the sense of cause in fact, has been defined as a cause "which in natural and continuous sequence, unbroken by any efficient intervening cause, produces the result complained of and without which the result would not have occurred." Smith v. Sharp, 82 Idaho 420, 426, 354 P.2d 172, 175 (1960). See also Chatterton v. Pocatello Post, 70 Idaho 480, 223 P.2d 389 (1950); Pilmer v. Boise Traction Co., 14 Idaho 327, 94 P. 432 (1908). This component of proximate cause embraces two closely related elements. First, an event is the cause in fact of a succeeding event only if the succeeding event would not have occurred "but for" the prior event. Thus, an act or omission is not the cause in fact of ensuing damage if the damage likely would have occurred anyway. The second element is a requirement that the first event be a "substantial factor" in producing the succeeding event. Munson v. State Department of Highways, 96 Idaho 529, 531 P.2d 1174 (1975). Thus, a defendant's conduct is the cause in fact of an event only if it was a material element and a substantial factor in bringing it about. W. PROSSER, supra, at 240; RESTATEMENT (SECOND) OF TORTS § 431 (1965).

In the present case, the jury was given an instruction on proximate cause corresponding to Idaho Jury Instructions (IDJI) No. 230. This instruction focuses upon cause in fact, but it does not expressly set forth either the "but for" element or the "substantial factor" element. Rather, the instruction simply explains "proximate cause" as:

a cause which, in natural or probable sequence, produced the damage complained of. It need not be the only cause. It is sufficient if it concurs with some other cause acting at the same time, which in combination with it, causes damage.

This instruction does not completely explain the elements of cause in fact. Moreover, the instruction does not make it entirely clear that in order for a concurring cause to give rise to liability, it must be a proximate cause in its own right. See Miller v. Northern Pacific Railway Co., 24 Idaho 567, 135 P. 845 (1913), and cases cited in 57 AM.JUR.2D Negligence §§ 176-82 (1971). However, this instruction was given without objection by either party. No issue concerning any of the jury instructions has been raised in this appeal. Consequently, we will confine our analysis to causation as the jury was instructed. The question put to the jury was whether the increased water depth in the canal, attributable to clogging at the trash rack, "produced" the damage "in natural and probable sequence ... in combination with" the increased flow of water entering the canal from the river.

A jury is entitled, when making a finding of causation, to draw an inference from circumstantial evidence. Henderson v. Cominco American, Inc., 95 Idaho 690, 518 P.2d 873 (1974). In general, a court should not substitute its determination of cause in fact for that of a jury, unless the evidence points only to one reasonable conclusion. Munson v. State Department of Highways, supra. Our Supreme Court has carefully defined the limits of reasonableness within which a jury can draw inferences from circumstantial evidence. In Splinter v. City of Nampa, 74 Idaho 1, 11, 256 P.2d 215, 221 (1953) — a case quoted with approval in Henderson — the Court said:

The underlying principle applicable here is that a verdict cannot rest on conjecture; that where a party seeks to establish a liability by circumstantial evidence, he must establish circumstances of such nature and so related to each other that his theory of liability is the more reasonable conclusion to be drawn therefrom; and that where the proven facts are equally consistent with the absence, as with the existence, of negligence on the part of the defendant, the plaintiff has not carried the burden of proof and cannot recover.

As noted previously, this case was presented to the jury not upon a theory of negligence but upon the State's violation of a contractual duty to avoid obstructions to the flow of water in the canal. However, the admonition in Splinter, relating to proof of negligence, is applicable to proof of any factual requisite of liability, by use of circumstantial evidence. Thus, in Dent v. Hardware Mutual Casualty Co., 86 Idaho 427, 435, 388 P.2d 89, 93 (1964), our Supreme Court — quoting with approval from Searles v. Manhattan Railway Co., 101 N.Y. 661, 5 N.E. 66 (1886) — stated:

When the fact is that the damages claimed in an action were occasioned by one of two causes, for one of which the defendant is responsible, and for the other of which it is not responsible, the plaintiff must fail if his evidence does not show that the damages were produced by the former cause, and he must fail, also, if it is just as probable that they were caused by the one as by the other, as the plaintiff is bound to make out his case by the preponderance of evidence. The jury must not be left to mere conjecture, and a bare possibility ... is not sufficient.

Dent does not stand on all fours with the instant case because the issue here is one of concurring causes rather than of independent causes. However, the fundamental lesson of Dent — that a plaintiff relying upon circumstantial evidence must show his inference of causation to be more probable than any opposing inference — is applicable here. Thus, the question on appeal is whether the jury, following the instructions given in this case, reasonably could have inferred that the canal failure more probably resulted from clogging at the trash rack, in combination with the increased flow from the river, than from such increased flow alone.

The jury was entitled to believe, and therefore we accept, the opinion of the canal company's expert that clogging of the trash rack could have contributed as much as three inches to the abnormal depth of water in the canal at the break point. These three inches represented a significant fraction (twenty percent) of the fifteen inches — over the normal water level in the canal — necessary to reach the top of the banks. The jury also was entitled to believe, and we similarly accept, other testimony that the rising river had "stabilized" by June 16 and that water then entering the canal was less than the quantities observed on other occasions. From this evidence we believe a jury permissibly could have inferred that the flow entering the canal from the river during the night of June 15-16 remained at a level within the canal's capacity; but that when this level was augmented by water backing up from the clogged trash rack, the canal failure followed in natural, probable sequence. We further believe a jury reasonably could have found this scenario more probable than the State's theory that the canal failed solely because of high water intake from the river.

The question admittedly is a close one. As our colleague notes in his dissenting opinion, the canal company's own expert testified that high water intake was the "primary" cause of the washout. However, such testimony did not shield the State from liability under the jury instruction on proximate cause given by the district court. We also acknowledge that the expert referred to the canal company's theory as "speculation." However, when examined in context of the expert's full testimony, this comment need not have been interpreted by the jury as an expression of disbelief in the company's claim. The same witness declared that "[t]he trash in the canal undoubtedly aggravated the situation." Finally, the expert conceded that when he calculated the increased depth due to clogging, he had not been informed of the elevation difference between the tops of the bypass gates and the water level reportedly reached at the trash rack. However, his testimony — again, taken as a whole — plainly indicated that his calculations were postulated upon an impedance of water flow at the trash rack, not upon a complete blockage. The fact that some water would have escaped into the bypass channel was not fatal to the expert's analysis.

We conclude that the jury's inference of causation was supported by substantial, though conflicting, evidence. The verdicts will not be set aside. Judgments upon the verdicts are affirmed. Costs to the canal company and farmers. No attorney fees on appeal.

WALTERS, C.J., concurs.

SWANSTROM, Judge, dissenting.

I do not agree that the jury verdicts in this case were supported by substantial evidence. Therefore, I dissent.

The following are undisputed facts which the jury could not ignore. During the time when the canal bank failed, the Salmon River was at a record flow and water from the river was pouring over the top and side of the headgate structure into the canal. The height of this water above the structure was at least fifteen inches, according to plaintiffs' own witnesses. The break occurred in a low spot in the canal bank, about 5,000 feet down from the headgate. The trash rack maintained by the state was another 4,500 feet below the break point.

The circumstantial evidence most favorable to the canal company was that, due to partial clogging of the trash rack from debris and due also to the excessive amount of water coming into the canal, the water level at the trash rack rose twenty-one inches, from its normal depth of 3.6 feet to a depth of 5.3 feet. It is important to remember that 5.3 feet was the maximum height of the water at the trash rack, according to the plaintiffs' evidence. According to evidence offered by defendant, it could not have reached that depth, but for purposes of this discussion I will assume it did.

Because of the slight downward slope of the canal the relative elevation of the canal floor at the break point was approximately four feet higher than the relative elevation of the canal floor at the trash rack. In other words, the average drop in the canal between the break point and the trash rack was about one foot in each 1,000 feet. The normal depth of the water at the trash rack was 3.6 feet, as noted above, while the normal depth at the point of the break was approximately 2.7 feet. In terms of relative elevation, the surface of the water at the point of the break was about 3.1 feet higher than at the trash rack. Therefore, if the water surface did rise a maximum of 1.7 feet, to 5.3 feet at the trash rack, the relative elevation of the surface of the water at the trash rack was still more than one foot lower than the normal relative elevation of the water surface where the break occurred. The actual difference, of course, was even greater than that because the surface of the water at the break point was above normal due to the Salmon River flooding into the upper end of the canal.

The critical question thus becomes: If clogging contributed to the rising of the water surface at the trash rack to 5.3 feet, how could that affect the water surface some 4,500 feet back up the canal when the surface there was already at a higher relative elevation? Plaintiffs attempted to show how this could have happened through the testimony of Dr. Watters, an expert on open channel hydraulics. He explained that, in a flowing open channel with a constant inflow, an obstruction at the downstream end, which diminishes the outflow, would create a "backwater curve" — in simple terms, a slowing of the stream with a resultant increase in depth of the flowing water in back of the obstruction for some distance. The amount of increase in depth is smaller as you go further back up the channel from the obstruction until it reaches a point at which the obstruction has no effect on the depth of the water. He explained that the increased depth of water, if any, at a given point in the open channel can be calculated using a complicated formula. The formula requires "plugging in" several "known" factors, such as the size, shape and slope of the canal, as well as the rate of flow of water into the canal and the depth of the water at the point of obstruction. The presence and effect of this backwater curve was crucial to plaintiff's case and was totally dependent upon Dr. Watters' testimony.

Dr. Watters based his testimony on both his own personal observations and the reports of others. Dr. Watters first visited the site four years after the washout. He inspected the canal and then made his calculations, using his own measurements and observations as to the size, shape, roughness and slope of the canal. He also relied upon some surveying data furnished by the state's expert and another engineer. However, in making his calculations he was forced to "assume" two of the required factors. He assumed that the water depth at the trash rack had reached 5.3 feet before the washout occurred. He also made certain assumptions as to the rate of flow, although he admitted he did not know what the flow actually was when the washout occurred and no other witness provided this information. Dr. Watters further testified that for the purpose of his calculations the rate of flow in the canal and the depth of water at the trash rack were "two critical pieces of information." He then estimated that if clogging had occurred as plaintiffs' contended, and if the flow rate was 175 cfs, it could have caused an increase in the depth of water at the place where the break occurred of from two to three inches. Dr. Watters, however, did not purport to say how deep the water was, without the backwater curve effect, at the point in the canal where the break occurred. The majority nevertheless says that this three-inch increase in water depth "represented a significant fraction (twenty percent) of the fifteen inches — over the normal water level in the canal — necessary to reach the top of the banks." At the place where the canal broke it would take forty-eight inches of water to fill the canal. Giving the canal company the benefit of all inferences, the state can only be held responsible for, at most, three inches of water depth out of a total water depth of at least forty-eight inches — an insignificant six percent.

Furthermore, Dr. Watters did not testify that clogging was the probable cause of the canal washout. On the contrary, he said:

I guess we really don't know just exactly how high the water had to get. I suppose it's possible that the trash rack clogging was just enough to precipitate the failure. A speculation, certainly, but something to consider.

In an earlier written report, introduced into evidence, Dr. Watters had stated that

it seems more justifiable to claim that the washout was caused primarily by high discharges,^[1] rather than trash rack clogging. The trash in the canal undoubtedly aggravated the situation, but should more appropriately be used as evidence of substantial overtopping of the canal headworks and a resulting excessive flow entering the canal. [Emphasis original.]

There was, in fact, substantial evidence that the excessive flow from the Salmon River into the canal alone could have overtopped and washed out the canal bank. Moreover, Dr. Watters was asked by plaintiffs' counsel if he thought the canal could have washed out even without the water actually overtopping the bank. Dr. Waters responded:

A. Oh. I think it's a possibility and one of the basis of the opinion is that the material in the area as I observed it during my visit was a very sandy, gravely type of material which I wouldn't advise using for canal banks but in any case when the water gets close to the top it will begin, and the bank gets narrower, then there will be more of a tendency for the flow to seep through the gravelly mixture and as the flow begins to move through the mixture it tends to washout grains and sand and create larger openings and sort of like if you build a basin or container out of sand and fill it up towards the top, as the water gets towards the top it begins to washout and eventually will washout a piece and as that piece washes out the whole thing comes apart. I think there's a possibility that could have been a contributing factor.

In making his calculations, Dr. Watters assumed a water depth of 5.3 feet at the trash rack due to clogging. However, he had not been informed that the tops of the gates in the bypass channel were only 4.3 feet high, relative elevation. He was therefore unaware that, before the water could reach 5.3 feet, it would flow over the bypass gates, through a twenty-four foot opening, and into the bypass channel. This is true even if the float-actuated gates remained shut. The majority lightly dismisses the importance of this undisputed evidence by saying "[t]he fact that some water would have escaped into the bypass channel was not fatal to the expert's analysis." Dr. Watters testified, however, that the depth of water at the trash rack was a "critical" piece of information for the purpose of his calculations. The evidence clearly showed it was highly improbable that the water depth could reach 5.3 feet because of the escapement structure the state had provided at the trash rack.

Finally, as to the effect of the debris itself, the evidence showed that it would raise the water level at the trash rack at most a few inches. Giving full credit to Dr. Watters' testimony about a backwater curve, the debris also would have produced only a minute rise in the water level where the canal broke — if it had any calculable effect at all. The escapement provided over the tops of bypass gates would further reduce, if not completely eliminate, the effect of the clogging if the water level in the canal rose above 4.3 feet at the trash rack. These considerations and a thorough review of the record leave me strongly convinced that there is a lack of substantial evidence to support the jury's verdicts in this case despite the "permissible inferences" the jury could have drawn. I would reverse the judgments.

UPON REHEARING

BURNETT, Judge.

After we issued our opinion affirming the jury verdicts, the Idaho Department of Fish and Game petitioned for rehearing. We granted the petition. Counsel for both sides reargued the case and ably responded to numerous questions from the Court. Having examined the factually complex record again, we remain persuaded that the jury verdicts should be upheld.

I

Our analysis rests upon two foundational points. The first is the standard of appellate review. The State would have us hold that the canal company failed to prove it more likely that the canal break resulted from the combined effects of high water and clogging at the trash rack than from the effect of high water alone. We believe the State has confused the canal company's burden of proof at trial with the proper standard of review on appeal. It is true that a plaintiff must show his damage to be the more probable result of a cause for which the defendant is responsible than of a cause for which the defendant is not responsible. Indeed, our principal opinion cites cases so holding. But we do not read those cases to vest in an appellate court the power to decide where the greater weight of probability falls. That remains for a jury to determine. Our function is to decide whether a jury reasonably could have found as did the jury in this case. We will not disturb jury verdicts supported by substantial and competent evidence, even though we entertain doubt as to which party's version of the conflicting evidence is more probable.

The second critical point is the jury instruction given in this case on proximate cause. Our principal opinion explains how this instruction, to which no one objected at trial, imposed no requirement upon the canal company to prove either that clogging at the trash rack was a "substantial factor" in causing the canal break or that the break would not have occurred "but for" the clogging. Under the instruction given, the canal company was required only to show that the clogging, "in natural and probable sequence ... in combination with" the water flow, caused the canal break. (Emphasis added.) Although we have criticized the jury instruction, it embodies a standard of causation upon which the parties mutually agreed to try the case.

II

Bearing in mind the standard for appellate review of jury findings, and the standard of causation submitted to the jury, we turn again to the evidence. The canal company's expert in open channel hydraulics testified that if the water depth in the canal had reached 5.3 feet at the trash rack, and if the flow of water in the canal had been approximately 175 c.f.s., the combined effects of the flow and of clogging at the trash rack would have produced an overtopping of the canal bank where the washout occurred. The witness further testified that such clogging would have contributed, in varying degrees, to the upstream water depths at other flow rates as well.

The State has not disputed the scientific methodology underlying this testimony. Rather, the State contends on rehearing that the expert's opinion was entitled to no weight because it was premised upon facts not in evidence. We disagree — and we note in passing that the State's capable trial counsel, who has not argued on appeal, never objected on this basis to the witness' testimony.

The record contains evidence of water depth and flow. As mentioned in our principal opinion, the depth of 5.3 feet at the trash rack was measured by a witness who observed high water marks the morning after the break. His testimony was reasonably consistent with that of another witness who measured high water marks upstream. We are mindful that these measurements seemingly conflicted with the fact that, of the four float-activated gates in the bypass channel, only one was found open after the washout. But we cannot say that the jury was compelled to disbelieve the witnesses' testimony concerning high water marks. The credibility of witnesses and the weight to be given their testimony are matters exclusively within the province of the jury. E.g., Harper v. Johannesen, 84 Idaho 278, 371 P.2d 842 (1962). A jury is entitled to disregard evidence only if it is inherently improbable. Our Supreme Court has held that "to warrant such action there must exist either a physical impossibility of the evidence being true, or its falsity must be apparent, without any resort to inferences or deductions." Olsen v. Hawkins, 90 Idaho 28, 37, 408 P.2d 462, 467 (1965), quoting Arundel v. Turk, 16 Cal. App. 2d 293, 60 P.2d 486, 488 (1936).

In this case, the observations of high water marks were not physically impossible. The bypass gates had been visually inspected, but not actually tested, shortly before the washout occurred. Moreover, the fact that the gates were 4.3 feet high does not conclusively establish — rather, it merely invites an inference — that the water in the canal never reached a depth of 5.3 feet at the trash rack. However, the jury also could have inferred that the combined effects of high water flow and clogging at the trash rack were sufficient to produce a depth of 5.3 feet despite a partial escapement of water over the bypass gates.

The record also contains evidence to support the postulated flow of approximately 175 c.f.s. The principal opinion notes that water from the Salmon River entered the canal by flowing over and around a debris-filled headgate. Witnesses who observed this flow before the washout testified that the water was some eight to twenty inches higher than a "wing wall" extending from the headgate into the river. A written report by the hydraulics expert, admitted into evidence to supplement his testimony, contained an undisputed statement that a twelve-inch flow over the wing wall would have produced a flow of approximately 175 c.f.s. in the canal. The expert also testified, as mentioned above, that even if differing flow rates were postulated, clogging at the trash rack would have contributed in varying degrees to the water depth at the point of the break. We hold that the expert's opinion was based upon facts in evidence, albeit conflicting in some respects, and was entitled to whatever weight the jury chose to give it.

III

Finally, the State argues on rehearing that the verdicts were products of an impermissible "stacking" of inferences. We agree that the jury stacked inferences. Implicitly, the jury inferred from circumstantial evidence that debris clogging the trash rack increased the water depth in the canal and further inferred that this increased depth, when combined with the flow entering the canal, caused the washout. However, we do not agree with the State that such stacking of inferences was impermissible. Again we note, in passing, that the State's capable trial counsel requested no special jury instruction on this point.

The State now invites attention to Mann v. Safeway Stores, Inc., 95 Idaho 732, 518 P.2d 1194 (1974). In that case our Supreme Court said, "[I]f an inference drawn from circumstantial evidence is to be made the basis of further inference, the first inference must be established to the exclusion of all other reasonable inferences." Id. at 737, 518 P.2d at 1199. Mann limits, although it does not prohibit, the stacking of inferences. The rigid application of such a limit to civil cases has come under critical scrutiny. See generally 1A J. WIGMORE, EVIDENCE IN TRIALS AT COMMON LAW § 41 (Tillers Rev. 1983). One of the anchor authorities for the Mann rule — and one cited in Mann itself — has been an Arizona Supreme Court decision in New York Life Insurance Co. v. McNeely, 52 Ariz. 181, 79 P.2d 948 (1938). But after McNeely was cited in Mann, the Arizona Court changed direction. In Farm-Aero Service, Inc. v. Henning Produce, Inc., 23 Ariz. App. 239, 532 P.2d 181, 184 (1975), the Court acknowledged that the rule announced in McNeely had created "much criticism and confusion." The Court stated that even in cases where McNeely had been applied, the courts actually were "concerned with the probative value of the evidence as a whole." Id.

Nevertheless, even if the Mann-McNeely rule were applied literally to the instant case, we are not convinced that the jury acted improperly. The evidence is undisputed that State employees removing debris before the washout observed the water level above the trash rack to be several inches higher than the level below the trash rack. They further noticed that the differential appeared to diminish as debris was removed. It is also undisputed that the employees departed from the site before all debris had been removed and while more debris was floating down the canal. Finally, it is undisputed that during the evening immediately prior to the washout, a farmer downstream from the trash rack observed the water level at his measuring device to be lower than normal. We think the only reasonable inference to be drawn from these undisputed facts is that debris clogged the trash rack to such an extent that it impeded the flow of water, increasing the depth above the trash rack and lowering the depth on the downstream side. Therefore, we find no error, under the Mann-McNeely rule, in the jury's use of this inference to support a second inference that the increased upstream depth, when combined with the flow entering the canal, produced the washout.

For these reasons we conclude, again, that the jury verdicts should be affirmed.

WALTERS, C.J., concurs.

SWANSTROM, J., adheres to the views expressed in his prior, dissenting opinion.

NOTES

[1] In using the term "high discharges" Dr. Watters was referring to the excessive water flowing from the Salmon River over the top of the headgate structure into the canal.