After paying a property damage claim arising from an airplane's crash into the home of its insured, Travelers Insurance Company (Travelers) brought this diversity action, as subrogee, against the owner of the plane and the estate of its deceased pilot. A jury found for the defendants, and Travelers appealed, assigning as error the district court's order compelling its substitution for the insured as the named plaintiff; the court's refusal, following Virginia law, to give a res ipsa loquitur instruction; its refusal to find the pilot negligent per se when he arguably violated a federal regulation requiring planes to fly above a certain height; and various evidentiary rulings. Finding no prejudicial error, we affirm the judgment.

* In November 1977, Joseph Reid was piloting an airplane owned by Harry Riggs from Columbus, Ohio to the Washington, D.C. area. After he had missed two approaches to the Dulles International Airport, the control tower gave him a heading to return toward Columbus. Soon after, the plane crashed into a townhouse owned by John and Veronica Frankenstein and insured by Travelers. Reid died as a result. Travelers paid the Frankensteins $26,997.99 for their loss, under a $100 deductible homeowners policy. This subrogation action was then commenced against Riggs and Reid's estate in the name of the Frankensteins as plaintiffs by counsel retained by Travelers. The plaintiffs sought by their prayer for relief to recover the sum of $26,997.99, the exact amount paid by Travelers.

During pretrial maneuvers, the district court, over Travelers' objection, granted the defendants' motion under Rule 17, Fed.R.Civ.P., to substitute Travelers for its insureds as sole plaintiff on the basis that it was the real party in interest by virtue of its total subrogation. Travelers then moved, two weeks before trial, to add the Frankensteins as plaintiffs and to increase the amount claimed by $710.60 (the amount of the deductible plus the difference between the replacement cost and the amount paid by Travelers for a damaged rug). Responding to Travelers' motion, the court suggested two options: (1) Travelers could sue for the total amount in its own name for the use and benefit of the Frankensteins, or (2) the Frankensteins could intervene in their own right. Travelers declined to pursue either option and stood on its objection to its substitution as sole party plaintiff.

At the ensuing jury trial the sole issue litigated was whether the crash was caused by the pilot's or owner's negligence or by some other cause. It was a typical battle of the experts; both sides offered the testimony of aviation experts who opined as to possible causes of the crash. Travelers also offered the testimony of eyewitnesses who suggested that the plane may have been flying below the minimum level set by federal regulations. The jury found for the defendants and the court refused to grant Travelers' motion for judgment n.o.v. This appeal followed.

* Travelers first contends that the district court erred in requiring the substitution of Travelers as sole named plaintiff in the action. For reasons that follow, we hold that in the end, no prejudicial error resulted from the district court's action.

It is apparent that the court originally treated the case as one involving total subrogation because the original claim was only for the amount paid by Travelers and did not include the deductible amount under the policy. On this basis the court must have considered that, under general subrogation and real party in interest principles, Travelers was the only real party in interest. It was apparently on this perception that the challenged substitution ruling was made. This ruling was technically in error because the proper focus of inquiry in application of these principles is not upon the amount claimed in the ad damnum clause but upon whether the insured has any uncompensated claim for which it may seek recovery. If the insured does have such a claim, it is a real party in interest in whose sole name the action may be prosecuted under general principles of subrogation. See, e.g., United States v. Aetna Casualty & Surety Co., 338 U.S. 366, 381, 70 S. Ct. 207, 215, 94 L. Ed. 171 (1949). Here, the Frankensteins as insureds did have such a claim-for the uncompensated deductible amount-and were accordingly, under general principles, real parties in interest who could not properly be displaced as parties plaintiff by the substitution of their partially subrogated insurer as sole named plaintiff.

We conclude that any prejudice from this technical error was cured by the court's subsequent action giving Travelers the option of having its insureds joined with it as co-plaintiffs by intervention. The ultimate effect of this option, if taken, would have been to force Travelers' joinder with its insureds as co-plaintiffs. Under the law of this circuit, that was the limit of its entitlement as a partially subrogated insurer. For, though not universally followed in the federal courts,1 and though subject indeed to criticism in this circuit,2 it is nevertheless the rule of this circuit that in a partial subrogation situation where the action is brought in the name of the insured as real party in interest, the partially subrogated insurer may also, if feasible, be involuntarily joined as a party plaintiff upon motion of the defendant. See Virginia Electric & Power Co. v. Westinghouse Electric Corp., 485 F.2d 78 (4th Cir. 1973); Ward v. Franklin Equipment Co., 50 F.R.D. 93 (E.D.Va.1970); Pinewood Gin Co. v. Carolina Power & Light Co., 41 F.R.D. 221 (D.S.C.1966). Since, in effect, the district court offered Travelers an option at least this favorable to its perceived interest, we cannot find prejudicial error in its overall action. That Travelers declined to take what was offered cannot of course be charged as error to the court.

Travelers contends, however, that under Erie and its progeny, this general circuit rule of involuntary joinder cannot be applied in this diversity case because of directly conflicting provisions of Virginia statutory law. Specifically, Va.Code § 38.1-31.2 (1981) allows an insurer who has either fully or partially paid an insured loss to sue either in its own name or that of its insured; and Va.Code § 8.01-5(B) (1977) forbids the involuntary joinder of an insurer that has elected under the former provision to sue in the name of its insured.

We do not think that this state rule prohibiting involuntary joinder of a subrogated insurer is binding in this federal diversity litigation. Even if the rule followed in this circuit is not one derived directly from Fed.R.Civ.P. 19, so that, under Hanna v. Plumer, 380 U.S. 460, 85 S. Ct. 1136, 14 L. Ed. 2d 8 (1965), it must be applied, we nevertheless think that faced with a "relatively unguided Erie choice," id. at 471, 85 S.Ct. at 1144, the proper choice is the federal rule.

Rules governing the joinder of named parties in litigation in no way affect "the primary conduct and affairs of (a state's) citizens," id. at 476, 85 S.Ct. at 1146 (Harlan, J., concurring). The general rule has therefore been that questions of joinder in diversity cases are appropriately determined by federal law. Provident Tradesmens Bank & Trust Co. v. Patterson, 390 U.S. 102, 125 n.22, 88 S. Ct. 733, 746 n.22, 19 L. Ed. 2d 936 (1968). The policy concern that undoubtedly underlies the Virginia rule-to avoid the prejudice that in conventional wisdom is assumed to afflict all insurance companies in jury trials-is clearly a legitimate one. But it has only to do with the conduct of litigation, and the resulting rule cannot be considered outcome-determinative in the sense of Guaranty Trust Co. v. York, 326 U.S. 99, 65 S. Ct. 1464, 89 L. Ed. 2079 (1945). Since choice of the federal rule is not in this sense outcome-determinative, any encouragement it lends to the limited forum shopping that is possible for defendants is simply of the kind that attends any federal rule of procedure thought by litigants to provide tactical advantage. Erie and its progeny have never required choice of the state rule to avoid federal forum choices on that basis.

Finally, in making an Erie choice of law not dictated by a specific federal rule of procedure, there must be weighed against the policies underlying a state rule asserted as controlling under Erie any countervailing federal policies respecting the litigation process. Byrd v. Blue Ridge Rural Electric Cooperative, 356 U.S. 525, 78 S. Ct. 893, 2 L. Ed. 2d 953 (1958). Here there are such policies, preeminently the policy that in federal litigation "(t)he pleadings should be made to reveal and assert the actual interest of the plaintiff, and to indicate the interests of any others in the claim." United States v. Aetna Casualty & Surety Co., 338 U.S. 366, 382, 70 S. Ct. 207, 216, 94 L. Ed. 171 (1949). This bespeaks an approach which abjures concealed or fictional issues and parties in interest and attempts to avoid unfair prejudice by jury voir dire, cautionary instructions, and other corrective measures rather than by a concealment of interests that is likely to be ineffectual in any event. See, e.g., Public Service Co. v. Black & Veatch, 467 F.2d 1143 (10th Cir. 1972) (contract, requiring each party to carry insurance, admitted into evidence). Were ultimate choice made to turn here on the countervailing force of the federal policies, we think they easily suffice.

Travelers' final contention on this point is that Fed.R.Civ.P. 17 itself, by expressly incorporating state law, necessarily prohibits its involuntary joinder as subrogated insurer in this Virginia diversity action. The provision relied upon is the clause in Rule 17(a) that "a party authorized by statute may sue in his own name without joining with him the party for whose benefit the action is brought." Because Va.Code § 38.1-31.2 (1981) allows an action to be brought solely in the name of an insured person, Travelers contends that involuntary joinder of the insurer is forbidden by the federal rule itself, independently of the prohibition contained in state law, Va.Code § 8.01-5(B) (1977).

We reject this contention. The provision cited simply requires federal courts to look to state law for designation of those persons who, as real parties in interest, may properly sue in their own names on state claims in federal courts. Virginia law does indeed provide that insureds, whether wholly or only partially compensated for insured losses, may sue in their names without joinder of their insurers as co-plaintiffs. But this provision of Rule 17(a) is properly interpreted only as defining-either directly or by incorporation of state law-those persons who, as real parties in interest, have substantive rights of action, and not as enabling those parties then to avoid joinder of other parties in interest. See Carlson v. Consumers Power Co., 164 F. Supp. 692 (W.D.Mich.1957); 6 C. Wright & A. Miller, Federal Practice and Procedure: Civil § 1543, at 645 (1971). Indeed, in the same context other federal courts have found no prohibition against the joinder or substitution of insurers despite the existence of state statutes which, like Virginia's, allow subrogated claims to be pursued in the name of the insured alone. See Rosenfeld v. Continental Building Operating Co., 135 F. Supp. 465, 469 (W.D.Mo.1955); White Hall Building Corp. v. Profexray Division of Litton Industries, Inc., 387 F. Supp. 1202, 1206-07 (E.D.Pa.1974), aff'd mem., 578 F.2d 1377 (3d Cir. 1978). We so hold here.

On this basis the district court did not err in giving Travelers the option of either suing in its own name as sole plaintiff or as co-plaintiff with its partially compensated insureds.

Travelers next challenges the district court's refusal to instruct the jury on res ipsa loquitur. To determine whether this refusal was error, we must examine Virginia law, for it is virtually uncontroverted3 that federal courts in diversity follow state law in deciding whether to apply res ipsa. See e.g., Lachman v. Pennsylvania Greyhound Lines, 160 F.2d 496 (4th Cir. 1947) (applying res ipsa in bus crash under Maryland law); 1A Moore's Federal Practice P 0.315(2), at 3204 (2d ed. 1981). See also 9 C. Wright & A. Miller, Federal Practice and Procedure: Civil § 2409 (1971) (analyzing the old Fed.R.Civ.P. 43). Fed.R.Evid. 302 now makes this general Erie-doctrine rule even clearer. See 21 C. Wright & K. Graham, Federal Practice & Procedure: Evidence § 5124 (1977). While federal courts have followed state law in applying res ipsa to airplane crashes, United Air Lines v. Wiener, 335 F.2d 379 (9th Cir. 1964), they may not apply res ipsa if the state court would not. See Detroit Edison Co. v. Knowles, 152 F.2d 422 (6th Cir. 1945); 1A Moore's Federal Practice P 0.315(2), at 3207.

On our reading of relevant decisions of Virginia's highest court, we conclude that under Virginia law, res ipsa loquitur is not applied in airplane crash cases. In Surface v. Johnson, 215 Va. 777, 214 S.E.2d 152 (1975), the Virginia Supreme Court said:

The plaintiff argues that we should apply the doctrine of res ipsa loquitur under the facts disclosed by the evidence here. Because the crash occurred, she would have us infer negligence on the part of Johnson, and then infer that his negligence was a proximate cause of the crash. We decline to draw such inferences from the mere happening of the crash, for it is a matter of common knowledge that an aircraft may fall or crash in the absence of negligence or fault on the part of its pilot.

214 S.E.2d at 154. See also Hall v. Payne, 189 Va. 140, 52 S.E.2d 76 (1949); Kelley v. Central National Bank, 345 F. Supp. 737 (E.D.Va.1972).

Travelers argues that we may ignore the various intimations against res ipsa's application in these decisions, particularly those in Surface, because they were all uttered as dicta. We do not think we can properly do so under our obligation to seek and apply state law as the highest court of that state would most likely apply it. In Surface it is true that the quoted language is, strictly speaking, dictum, since the Surface court ultimately held that the plaintiff's proof sufficed without aid of res ipsa to make a submissible case. Certainly, the court's decision could stand without its apparent rejection of res ipsa. But the res ipsa point was specifically raised by the plaintiff, urged as a basis for decision, and rejected by the court. If the Virginia Supreme Court had thought res ipsa was a proper rule for airplane crash cases in Virginia, Surface was an appropriate time to state that dismissals are never proper if the plaintiff has shown the necessary predicates for res ipsa.

Travelers also seeks to distinguish the Virginia cases by noting that they all involve victims who were passengers on the doomed aircraft. The victims conceivably could have interfered with the pilot's operation of the plane, and thus the plaintiffs could not show a necessary ingredient for res ipsa-exclusive control by the defendant.4 It is true that the Hall and Kelley courts did base their rejection of res ipsa in part on this point. See Hall v. Payne, 52 S.E.2d at 79; Kelley v. Central National Bank, 345 F. Supp. at 741. But while the victim in Surface was indeed a passenger, the Virginia Supreme Court did not use this rationale in rejecting res ipsa. Thus, Surface stands for a broader proposition than Travelers suggests.

This is not a situation, then, where the state law is completely ambiguous or suspect from age. Indeed, our research reveals no cases even hinting that Virginia decisions are not unanimous on this point. Further, the denial of res ipsa in airplane crash cases, while perhaps no longer the majority view in the country, is not so unreasonable that we could confidently predict that the Virginia courts would go the other way if directly faced with the issue today. Because the best evidence we have is that Virginia would apply the reasoning of Surface if it were faced with our situation, we hold that the district court properly interpreted Virginia law in refusing to instruct the jury on res ipsa loquitur.5

We have also considered and find no basis for reversal in Travelers' assignments of error to several evidentiary rulings and the refusal of the district court to direct a verdict in its favor.

Travelers sought to introduce a 65-page report of the crash made by the National Transportation Safety Board (NTSB). The court admitted the factual portions of the report but refused to admit its conclusions. These conclusions were that an autopsy revealed "no preexisting physical or physiological condition that would have affected the pilot's performance of duties," and that an examination of the plane's engines revealed "no indication of pre-impact difficulty." Both conclusions increased the possibility that pilot negligence caused the crash.

Federal Rule of Evidence 803(8) suggests that, as a general matter, public reports, including evaluative sections, are admissible evidence. See Fed.R.Evid. 803(8), Notes of Advisory Comm. However, 49 U.S.C. § 1441(e) explicitly states that no part of any accident report of the National Transportation Safety Board shall be admitted as evidence in any action for damages. We need not examine whether the district court correctly interpreted § 1441(e) as allowing factual portions of NTSB reports to be admitted, because neither side disputes this here. We do think that, at the least, § 1441(e) forbids the use of conclusory sections of NTSB reports, and thus hold that the district court properly excluded them. Cf. Kline v. Martin, 345 F. Supp. 31 (E.D.Va.1972) (NTSB investigators not required to answer discovery deposition questions about possible pre-impact malfunction of the aircraft or incapacity of the pilot).

Travelers next argues that the trial court unfairly supervised the battle of the experts by requiring its expert to base his opinion solely on the record, while allowing the defendants' experts more latitude. Further, insists Travelers, the court improperly allowed Velta Benn, one of the defendants' experts, to base her opinion on the hypothesis that the pilot was not negligent. Finally, says Travelers, the trial court treated its expert more curtly than it did the defendants' experts, so that the court's overall handling of the matter was prejudicial even if the court's specific rulings were technically correct.

We find no abuse of trial court discretion in the respects charged. The trial court attempted to limit all expert testimony to opinions based on the NTSB report. Travelers' expert was allowed to present his opinion, based on the report's transcription of communications between the pilot and control tower, that pilot overload caused the crash. Benn suggested that the pilot's loss of consciousness or death, and subsequent slumping on the yoke, caused the crash. Such pressure on the yoke would cause, said Benn, a descent and crash similar to the one described in the NTSB report. She based her opinion in part on the fact-again, derived from the NTSB report-that the flight had been proceeding normally, but not on the assumption of no pilot negligence. The opinions of both experts were properly based on facts in the NTSB report, and it was the jury's function to decide between them. On the record we review we cannot find reversible error in the trial court's relative treatment of the opposing experts.

Travelers argues finally that the trial court erred in refusing to direct a verdict because the evidence conclusively established that the pilot was negligent per se in flying below the minimum height required by federal regulations, 14 C.F.R. § 91.79 (1981). According to Travelers, the evidence clearly showed that the plane broke the cloud cover and was skimming trees and houses shortly before it crashed. Even if this were true, it may well have been because the plane was already out of control when it reached this level, a possibility consistent with the expert testimony of Benn. If so, the pilot would not be negligent per se. In the context of this case, the issue was one for the jury, and the court properly refused to direct a verdict on this basis.

For the foregoing reasons, we find no error, and accordingly affirm the judgment of the trial court.

AFFIRMED.