Arreguin-Leon v. Hadco Construction

2020 UT 59, 472 P.3d 927

Opinions

Opinion

This opinion is subject to revision before final publication in the Pacific Reporter

2020 UT 59

IN THE

SUPREME COURT OF THE STATE OF UTAH

NOE ARREGUIN-LEON,

Petitioner,

v.

HADCO CONSTRUCTION, LLC,

Respondent.

No. 20190121

Heard February 10, 2020 Filed August 17, 2020

On Certiorari to the Utah Court of Appeals

Fourth District, Provo

The Honorable Fred D. Howard No. 130400816

Attorneys:

Troy L. Booher, Beth E. Kennedy, Salt Lake City, Leonard McGee, Peter Mifflin, Sandy, for petitioner Robert L. Janicki, Michael L. Ford, Sandy, Harry Lee, Shannen W. Coffin, Mark C. Savignac, Washington D.C., for respondent

JUSTICE PETERSEN authored the opinion of the Court, in which CHIEF JUSTICE DURRANT, ASSOCIATE CHIEF JUSTICE LEE, JUSTICE HIMONAS, and JUSTICE PEARCE joined.

JUSTICE PETERSEN, opinion of the Court:

INTRODUCTION

¶1 Noe Arreguin1 was injured while working on a highway construction site. He sued Hadco Construction, LLC, the general

__________________________________________________________

1 Although the plaintiff’s last name in the case caption is Arreguin-Leon, we refer to him in this opinion as Arreguin because that is how he refers to himself in his briefing.

Opinion of the Court

contractor, for failing to take necessary safety measures to protect workers from highway traffic. Arreguin prevailed at trial. But during trial, he elicited undisclosed testimony from his expert witness. The court of appeals found this error to be harmful and reversed and remanded for a new trial. We affirm.

BACKGROUND

¶2 Noe Arreguin was injured while installing an exit sign on the shoulder of I-15. A driver fell asleep at the wheel and veered off the road and into the ladder on which Arreguin was standing.

¶3 Arreguin worked for a company called Highway Striping & Signs. The company had been hired by Hadco to install signage for a Utah Department of Transportation roadway project in Utah County. In its role as general contractor, Hadco was responsible for implementing a “traffic control plan” composed of various safety measures to protect workers from traffic and drivers from the construction site. Hadco did not do so. At the time of the accident, there were no traffic control measures in place at the accident site, such as barrels or barriers.

¶4 As a result of the accident, Arreguin sustained significant injuries. He sued the driver and Hadco (along with others who are not relevant to this appeal).

¶5 Arreguin retained Bruce Reading as an expert witness on traffic control standards. Hadco’s counsel elected to depose Reading rather than receive an expert report.

¶6 The case proceeded to a jury trial, at which Arreguin called Reading to testify. Reading opined that Hadco or its subcontractor had violated five specific engineering practices, regulatory standards, and contractual provisions and that there was no traffic control plan in place at the accident site.

¶7 During direct examination, Arreguin’s counsel asked Reading, “If [200 yards from the construction project is] where [the driver] started to exit the roadway, what effect would a correctly installed buffer zone have had on his driving?” Hadco’s counsel objected and asked to approach the bench. The following sidebar ensued:

[DEFENSE COUNSEL]: Seems to me like this testimony is going toward causation—would traffic control have prevented the accident—and it goes beyond any opinion that he’s ever disclosed in this

2

case. There’s a list of his items of testimony, and he doesn’t touch on that at all. [PLAINTIFF COUNSEL]: Your Honor, Mr. Reading was deposed in this case. [Defense counsel] had every opportunity to ask any question he wanted, and—and he’s not limited to the initial disclosure. If he had—if [defense counsel] had elected a report, he would be limited to the contents of the report, but because a deposition has been elected, Mr. Reading is not so limited. [DEFENSE COUNSEL]: That’s not correct, your Honor. [PLAINTIFF COUNSEL]: And—and there were documents provided to Mr. Reading after (inaudible). [DEFENSE COUNSEL]: Then he needs to supplement his disclosure. THE COURT: Your objection is noted and is, frankly, overruled. [DEFENSE COUNSEL]: Can I make a record—a record on this? I think it’s very important. THE COURT: This record is the record here now. [DEFENSE COUNSEL]: Okay. Thank you.

¶8 Reading then testified about the effect that a proper traffic control plan would have had, including that if the accident occurred where Hadco’s “safety person”2 suggested it did, it would have been within a 900-foot area where the driver would have hit “at least one, if not more, of th[e] plastic barrels” that would have been in place. He explained that after hitting at least one of the plastic barrels, the driver “would have had close to six seconds to wake up and take corrective action.” And he concluded that if traffic control had been in place, “[t]here might have been an accident still,” but it would not have taken place where it did.

__________________________________________________________

2 At trial, Reading referred to Hadco’s “safety person.” From the context, we understand this person to be the Hadco employee who completed Hadco’s incident report about the accident.

3

¶9 The trial continued and the jury ultimately found that Hadco was partially liable for Arreguin’s injury. The jury allocated 60 percent of the fault to the driver and 40 percent to Hadco. Hadco appealed.

¶10 Approximately four months after filing its notice of appeal, Hadco filed Reading’s deposition transcript and Arreguin’s expert disclosures in the district court. They were included in the record that was certified to the court of appeals.

¶11 One of Hadco’s arguments on appeal was that the district court erred under Utah Rule of Civil Procedure 26 when it allowed Reading to offer an undisclosed opinion on causation. Arreguin argued that Hadco could not prevail on this argument without relying on his expert disclosures and Reading’s deposition transcript. But he asserted it was improper for the court of appeals to consider these documents because they were not actually part of the trial record.

¶12 The court of appeals decided to consider the disclosures and the deposition transcript. It “acknowledge[d] that [Reading’s] deposition was filed after the judgment was entered in this matter, but before the record was prepared” and that “such filings normally would not put the deposition before [the court of appeals] for consideration.“ Arreguin-Leon v. Hadco Constr. LLC, 2018 UT App 225, ¶ 6 n.2, 438 P.3d 25. But it decided that “under the unique facts of this case” it would “exercise [its] discretion and consider the deposition.” Id.

¶13 The court of appeals concluded that the district court abused its discretion in allowing Reading to testify about causation at trial. Id. ¶ 20. And it determined that the error was “harmful enough to warrant reversal and a new trial.” Id. ¶ 32.

¶14 Arreguin petitioned this court for certiorari, which we granted to address three questions: (1) “[w]hether the Court of Appeals erred in considering a deposition transcript that was not included in the record prior to the filing of the appeal”; (2) “[w]hether the Court of Appeals erred in construing [Hadco’s] arguments on appeal to present a sufficient basis for its conclusion that [Arreguin’s] expert testimony should have been excluded”; and (3) “[w]hether the Court of Appeals erred in its construction and application of the standard for demonstrating harmful error on appeal.”

¶15 We have jurisdiction pursuant to Utah Code section 78A-3-102(3)(a).

4

STANDARD OF REVIEW

¶16 “On certiorari, we review the court of appeals’ decision for correctness, without according any deference to its analysis.” Vander Veur v. Groove Entm’t Techs., 2019 UT 64, ¶ 7, 452 P.3d 1173 (citation omitted) (internal quotation marks omitted).

ANALYSIS

I. CONSIDERATION OF ARREGUIN’S EXPERT DISCLOSURES AND THE EXPERT’S DEPOSITION TRANSCRIPT

¶17 The first question before us is whether the court of appeals erred in considering Arreguin’s expert disclosures and Reading’s deposition transcript. Arreguin argues that the court of appeals should not have considered these documents because they were not truly part of the trial record, in that neither party submitted either document for the district court’s consideration at any point. Rather, Arreguin notes that Hadco filed the documents five months after the district court entered the final judgment in the case and four months after Hadco filed its notice of appeal.

¶18 We take Arreguin’s point. When the district court ruled on Hadco’s objection, it did not actually have these documents before it. The court of appeals decided to consider this extra-record evidence because it determined Hadco’s counsel had attempted to make a further record during the sidebar but had been prevented from doing so by the district court. See Arreguin-Leon v. Hadco Constr. LLC, 2018 UT App 225, ¶ 6 n.2, 438 P.3d 25. The court of appeals analogized this situation to one in which a party is prevented from objecting, and thereafter should not be prejudiced by the lack of an objection. See id.; see also UTAH R. CIV. P. 46. Arreguin rejects this as a valid basis for considering the extra-record materials.

¶19 We conclude that we do not need to resolve this dispute.

We can affirm the court of appeals’ ruling on the disputed expert testimony without resort to the expert disclosures or deposition transcript. Arreguin’s premise is that the content of these documents is essential to Hadco’s argument—in other words, that Hadco cannot successfully argue that the district court erred in permitting Reading to testify about causation without relying on the content of (1) the expert disclosures to prove that Arreguin did not in fact disclose a causation opinion and (2) the deposition transcript to prove that Hadco had “locked in” Reading to only those opinions he offered during his deposition. But as we will discuss, Arreguin never put the content of these documents at

5

issue in the district court. Arreguin essentially contends that Hadco must refute arguments he never made.

¶20 The court of appeals did rely upon the documents in its reasoning, Arreguin-Leon, 2018 UT App 225, ¶ 23. But we disagree with the premise that they are necessary to Hadco’s argument or the court of appeals’ holding. As Hadco has argued, we can resolve the rule 26 issue based solely on the transcript of the sidebar between counsel and the district court at trial.

¶21 Looking only at the sidebar, we agree with the court of appeals that the district court committed legal error in overruling Hadco’s objection. During the sidebar, Hadco argued that Arreguin’s question to Reading elicited a causation opinion, which went “beyond any opinion that [Arreguin had] ever disclosed in this case.” It is important to note Arreguin’s response. He did not dispute Hadco’s assertion that his question would elicit causation testimony. He did not assert that he had in fact disclosed that Reading would offer a causation opinion or that Reading had discussed causation in his deposition. And he did not argue that Hadco had failed to “lock in” Reading to only the opinions he had given at the deposition and therefore Reading was free to offer additional opinions.

¶22 Instead, Arreguin asserted broadly that the expert was not limited at all because Hadco had opted for a deposition rather than an expert report. Arreguin’s counsel stated, Your Honor, Mr. Reading was deposed in this case. [Defense counsel] had every opportunity to ask any question he wanted, and—and he’s not limited to the initial disclosure. If he had—if [defense counsel] had elected a report, he would be limited to the contents of the report, but because a deposition has been elected, Mr. Reading is not so limited.

Hadco’s counsel responded, “That’s not correct, your Honor.” With these arguments before it, the court overruled Hadco’s objection and permitted the expert to offer the disputed testimony.3

__________________________________________________________

3 After this back-and-forth but before the court ruled, Arreguin interjected that Hadco had provided documents to Reading after the deposition. And Hadco countered, “[t]hen he needs to (continued . . .)

6

¶23 As the court of appeals correctly concluded, the district court’s ruling was legally incorrect. Id. ¶ 26. Just because a party opponent selects a deposition rather than an expert report does not mean an expert’s subsequent trial testimony can be a “free-for-all.” Id. ¶ 21.

¶24 In general, rule 26 provides that “discovery may be obtained from an expert witness either by deposition or by written report.” UTAH R. CIV. P. 26(a)(4)(B). With respect to a written report, the rule makes clear that an expert is limited to opinions disclosed in the report. Id. (“A report shall . . . contain a complete statement of all opinions the expert will offer at trial and the basis and reasons for them. Such an expert may not testify in a party’s case-in-chief concerning any matter not fairly disclosed in the report.”). The rule itself does not make a similar statement with regard to an expert’s deposition. We generally agree with the relevant advisory committee note, which explains that “[i]f a party elects a deposition, rather than a report, it is up to the party to ask the necessary questions to ‘lock in’ the expert’s testimony.” Id. advisory committee note. In a case where an opposing party fails to “lock in” an expert witness during the deposition, the opposing party runs the risk of surprise testimony at trial.

¶25 However, this does not equate to the blanket assertion advanced at trial by Arreguin that if Hadco “had elected a report,

supplement his disclosure.” Hadco’s response was legally correct. See UTAH R. CIV. P. 26(d)(3)–(4); see also Arreguin-Leon v. Hadco Constr. LLC, 2018 UT App 225, ¶ 23, 438 P.3d 25. And we do not think that this additional exchange requires a review of the content of the disclosures. Here too, Arreguin did not respond that he had supplemented his disclosures or otherwise provided a causation opinion at some point before trial. So he did not put the content of his disclosures at issue.

Arreguin asserts in his briefing to us that Hadco forfeited its argument that Arreguin failed to supplement his disclosures. Arreguin argues that although Hadco preserved this argument during the sidebar, it did not raise the same argument in its briefing to the court of appeals, so the argument has been waived. However, Hadco has consistently asserted that Arreguin never disclosed or produced, in any form or at any time before trial, a causation opinion from Reading. Accordingly, we reject Arreguin’s preservation argument.

7

[Reading] would be limited to the contents of the report, but because a deposition has been elected, Mr. Reading is not so limited.” Arreguin’s counsel did not put the contents of the disclosures or deposition at issue by asserting that a causation opinion had in fact been disclosed at some point or that Hadco had not properly “locked in” Reading at his deposition and therefore he was free to offer new opinions at trial. Rather, Arreguin made a very broad assertion to the district court that if a party opponent elects to depose an expert witness, then the expert witness is not limited during trial testimony. This is an incorrect interpretation of rule 26(a)(4)(B), which the district court should have rejected.

¶26 Accordingly, we agree with the court of appeals that the district court should not have permitted Reading to offer the disputed testimony based on the arguments before it. See id. 26(d)(4) (“If a party fails to disclose or to supplement timely a disclosure or response to discovery, that party may not use the undisclosed witness, document or material at any hearing or trial unless the failure is harmless or the party shows good cause for the failure.”). And we conclude that although the court of appeals reviewed the disclosures and deposition and determined that Arreguin had not disclosed a causation opinion and that Hadco did in fact “lock in” Reading to the opinions he provided at the deposition, this was not necessary to reach the correct legal result because Arreguin had never argued otherwise.4 II. UTAH RULE OF APPELLATE PROCEDURE 11

¶27 Although we do not need to reach the issue of whether the court of appeals erred in considering the deposition transcript, the parties’ briefing and oral argument did elucidate certain ambiguities in Utah Rule of Appellate Procedure 11 that we flag for our appellate rules advisory committee. The relevant portions of rule 11 state, (a) Composition of the record on appeal. The original papers and exhibits filed in the trial court,

__________________________________________________________

4 Arreguin also argues in his briefing to us that “because Hadco did not make the documents part of the record, Hadco could not show prejudice on appeal.” We are unsure how to interpret this argument, and we are unable to resolve it because Arreguin does not explain the argument further.

8

. . . the transcript of proceedings, if any, the index prepared by the clerk of the trial court, and the docket sheet, shall constitute the record on appeal in all cases. . . . Only those papers prescribed under paragraph (d) of this rule shall be transmitted to the appellate court. ... (d)(2) Civil cases. Unless otherwise directed by the appellate court upon sua sponte motion or motion of a party, the clerk of the trial court shall include all of the papers in a civil case as part of the record on appeal. ... (h) Correction or modification of the record. If any difference arises as to whether the record truly discloses what occurred in the trial court, the difference shall be submitted to and settled by that court and the record made to conform to the truth. If anything material to either party is misstated or is omitted from the record by error, by accident, or because the appellant did not order a transcript of proceedings that the appellee needs to respond to issues raised in the Brief of Appellant, the parties by stipulation, the trial court, or the appellate court, either before or after the record is transmitted, may direct that the omission or misstatement be corrected and, if necessary, that a supplemental record be certified and transmitted.

UTAH R. APP. P. 11.

¶28 First, rule 11 references “the record” throughout, but it does not define it. Rule 11(a) states that the “original papers and exhibits filed in the trial court, . . . the transcript of proceedings, if any, the index prepared by the clerk of the trial court, and the docket sheet, shall constitute the record on appeal in all cases.” Id. 11(a) (emphasis added). It goes on to say that “only those papers prescribed under paragraph (d) . . . shall be transmitted to the appellate court.” Id. Rule 11(d)(2), which relates specifically to civil cases, is quite broad. It states that the clerk “shall include all of the papers in a civil case as part of the record.” Id. 11(d)(2). But as illuminated here, there is ambiguity as to what “all of the papers in a civil case” includes. Arreguin asserts that the record

9

should include only those items that were actually presented in court in some manner. However, the rule does not make explicit that this is the case.

¶29 Next, rule 11(h) provides a mechanism for parties to correct or modify the record. It states that, “[i]f any difference arises as to whether the record truly discloses what occurred in the trial court, the difference shall be submitted to and settled by that court and the record made to conform to the truth.” Id. 11(h). This has a clear meaning in some contexts—for example, if a transcript inaccurately documents a witness’s testimony and can be corrected by comparing the transcript to the audio recording of the testimony, that would seem to make the record conform to the truth. But in other contexts, the scope of what is meant by making the record “conform to the truth” may not be entirely clear.

¶30 Finally, rule (11)(h) also permits modification of the record “[i]f anything material to either party is misstated or is omitted from the record by error” or “by accident.” But the rule does not define either of these terms, and the scope of what might be encompassed within them is not entirely clear.5 III. HARMLESS ERROR

¶31 Next, we must determine whether the court of appeals erred in its construction and application of the standard for demonstrating harmful error on appeal. We conclude it did not.

__________________________________________________________

5 We also flag for our civil rules advisory committee a concern raised by Arreguin’s counsel at oral argument regarding the court of appeals’ treatment of Utah Rule of Civil Procedure 50(b). By way of background, Arreguin argued in the court of appeals that Hadco’s claims on appeal were unpreserved because Hadco had not renewed its motion for directed verdict after trial, and therefore had failed to meet the procedural requirements of rule 50(b). The court of appeals ultimately did not need to resolve this argument but briefly addressed it in a footnote. See Arreguin-Leon v. Hadco Constr. LLC, 2018 UT App 225, ¶ 29 n.9, 438 P.3d 25. At oral argument before this court, counsel for Arreguin raised a concern with this footnote. Because this issue is not before us, we do not address it or opine one way or the other on the court of appeals’ take on the rule. But we refer counsel’s concern to our civil rules advisory committee for consideration.

10

¶32 Arreguin argues that the court of appeals misapplied the harmlessness standard, citing to the court of appeals’ statement that it could not conclude “that the jury would have inevitably reached the same result without Expert’s testimony.” Arreguin-Leon v. Hadco Constr. LLC, 2018 UT App 225, ¶ 28, 438 P.3d 25. He contends that this is the incorrect standard.

¶33 Arreguin is correct that the cited sentence is not the correct harmlessness standard. Rather, “[h]armless error is an error that is sufficiently inconsequential that there is no reasonable likelihood that it affected the outcome of the proceedings.” H.U.F. v. W.P.W., 2009 UT 10, ¶ 44, 203 P.3d 943 (citation omitted). However, although the court of appeals used the language Arreguin identifies, it does not appear to us that it mistakenly thought this was the applicable legal standard. In the preceding paragraph, the court of appeals correctly stated that “[a]n error is harmful ‘only if the likelihood of a different outcome is sufficiently high as to undermine our confidence in the verdict.’” Arreguin-Leon, 2018 UT App 225, ¶ 27 (citation omitted). This is substantively similar to the legal standard we identify above. When the court used the disputed language, we agree with Hadco that it was merely a shorthand reference to the legal standard it had already identified. We see no legal error in the court of appeals’ application of the law.

¶34 Arreguin also argues that the court of appeals erred when it found the district court’s error to be harmful. He argues that “expert testimony is not required to establish [the] obvious proposition” that “if Hadco had set out the barrels to block traffic, the driver would have hit at least one of them, woken up, and taken corrective action” and “the barrels with sand in the bottom likely would have prevented the driver from crashing into [Arreguin].” Accordingly, he asserts that any error was harmless because Reading’s testimony was “unnecessary and cumulative of common sense” and “stated the obvious.”

¶35 We are not convinced. After the district court overruled Hadco’s objection, Reading went beyond his testimony about the components of a proper traffic safety plan and gave his opinion of how such a safety plan would have changed the events that led to Arreguin’s injury. Reading testified that if the driver drifted off the road 200 yards back, he would have hit a barrel and “would have been aware immediately upon impact” of the barrel. He estimated that based on a two-and-a-half second reaction time, the driver would have had “six seconds to wake up and take

11

corrective action.” He then would have “jerk[ed] hard left.” Ultimately, Reading opined that “[t]here might have been an accident still. There’s no question about that. I don't think the accident would have taken place where this happened.”

¶36 We think this testimony goes beyond common sense. A lay juror could be expected to understand the gist of Reading’s causation opinion—that the sleeping driver might hit a barrel, wake up, and attempt a correction. But a layperson would not necessarily understand with such precision the effect of a traffic safety plan upon the events in question. And even if a layperson might have assumed that “the accident would [not] have taken place where this happened,” this opinion carried extra weight because it came from an expert. We agree with the court of appeals’ observation that Reading’s testimony “carried the imprimatur of coming from an ‘expert,’” and it “provided a logical roadmap that the jury could—and likely did—follow in deciding the issues of liability and in apportioning fault.” Id. ¶ 28.

¶37 Arreguin also argues that the error was harmless because Reading’s testimony was cumulative of testimony given by other witnesses. But we do not view the testimony from the other witnesses to be equivalent to Reading’s causation testimony.

¶38 Arreguin argues that Reading’s testimony that the sleeping driver would have hit a barrel was duplicative of testimony from Hadco’s expert. But this is not so clear. Reading testified that given the parameters suggested by Hadco’s “safety person,” the driver “would have [hit][6] at least one, if not more, of these plastic barrels.” In comparison, Hadco’s expert testified about calculating tapers and spacing of traffic control devices, but he did not say that the sleeping driver would have hit a barrel. The closest Hadco’s expert came to saying this was noting that a traffic control device is an “indicator” and that barrels are “not going to stop a vehicle from departing the roadway,” while agreeing that they would “notify.” Reading’s testimony was more specific and certain, and it was not cumulative of the testimony from Hadco’s expert.

__________________________________________________________

6 The transcript says this word was inaudible. From the context of the sentence, it appears that the word was “hit” or another synonymous word.

12

¶39 Arreguin also argues that Reading’s testimony that hitting the barrel would have awakened the driver was cumulative of testimony from Hadco’s expert, Hadco’s project manager, and the driver. Reading opined that hitting a barrel would lead to a “hellacious sound” that is “going to wake him up.” In contrast, when asked “if striking a barrel can be a jolting, noisy experience,” Hadco’s expert responded, “Well, yes.” He then commented, “I can’t say, if someone is already asleep, though, if that would be something that would necessarily wake them up.” When asked, “Would you agree that it is possible that there was a barrel on the side of the road and [the driver] hit it as he was going off the road, that it may have alerted him,” Hadco’s expert responded, “It could have. I mean, I—I would be speculating, but yeah. I couldn’t say specifically that it would, but it may have.” Hadco’s project manager agreed that barrels need to be “crashworthy.” And the driver testified that he woke up when he heard “the grids in the road.” Arreguin argues that based on this testimony, the jury could infer that if a rumble strip awakened the driver, hitting a barrel also would have awakened him. While that might be a fair inference, we disagree with Arreguin’s assertion that Reading’s testimony is merely duplicative of the other witnesses’ testimony. Again, Reading’s testimony was specific and certain, while the testimony of the other witnesses was equivocal or required an inferential leap.

¶40 Arreguin next argues that Reading’s testimony that the driver would have taken corrective action also came from Hadco’s expert and the driver. Again, Hadco’s expert testified about calculating tapers and spacing of traffic control devices. He did not clearly state that the driver would have taken corrective action. Similarly, the driver testified that upon waking up and seeing a flatbed truck in front of him he “swerved off to the side to avoid it.” Neither is equivalent to Reading’s testimony that the driver “would have had close to six seconds to wake up and take corrective action,” that “the normal experience is you jerk hard left to get back on,” and finally that “[t]here might have been an accident still,” but he did not “think the accident would have taken place where this happened.”

¶41 Reading’s disputed testimony related to the important questions of whether and to what extent Hadco’s failure to implement a proper traffic control plan on the day of the accident caused Arreguin’s injuries. While other witnesses made statements from which the jury could possibly have inferred the disputed facts and opinions Reading provided, none of them gave

13

testimony that was equivalent to Reading’s. None of the other witnesses’ testimony on the disputed points was as clear, specific, and emphatic as Reading’s. We agree with the court of appeals that the district court’s error was not harmless. The erroneously admitted testimony was not “sufficiently inconsequential that there is no reasonable likelihood that it affected the outcome of the proceedings.” See H.U.F., 2009 UT 10, ¶ 44 (citation omitted).

CONCLUSION

¶42 We agree with the court of appeals that the district court abused its discretion in allowing Reading to offer causation testimony. This error was harmful. We affirm the court of appeals’ decision, and we remand to the district court for a new trial.

14