Melville v. American Home Assurance Co.

443 F.Supp. 1064 (1977)

Virginia J. Barry MELVILLE
v.
AMERICAN HOME ASSURANCE COMPANY.

Civ. A. No. 73-1398.

United States District Court, E. D. Pennsylvania.

November 25, 1977.

Gerald A. Robbie, New York City, for plaintiff.

Sidney L. Wickenhaver, Carol A. Mager, Philadelphia, Pa., for defendant.

INDEX
I. Preliminary Statement                        1068
II. The Procedural History                       1070
III. The Facts Adduced at Trial                   1070
IV. Choice of Law                                1076
A. Introduction                              1076
B. Is the Presumption Against Suicide
Procedural or Substantive?                1077
C. Must We Choose? — The Relevant
Similarities and Differences Among
the Laws of Pennsylvania, New York,
and Delaware                              1080
D. The Three Competing Choice of Law
Systems                                   1082
E. Pennsylvania Choice of Law Doctrine;
the Legacy of Griffith and its Impact
on Contract Cases                         1084
1. The Pennsylvania Supreme Court
and Superior Court Decisions            1085
2. The Court of Appeals (Third Circuit)
Decisions                              1090
3. A Disquieting Verdict                  1094
F. The Parties' Contentions; The Impact
of Van Dusen v. Barrack                   1094
G. Our Pretrial Bench Opinion                1097
H. Application of Pennsylvania Choice of
Law Principles to this Case               1097
1. Analysis Under Restatement I and
Lex Loci (Contractus) Rules            1097
2. Post-Griffith Interest/Restatement
II Hybrid Analysis                      1100
(a) Interest Analysis: Lex Loci v.
Lex Fori in the "Unprovided-for
Case."                              1100
(b) Territorial Contacts Analysis       1104
3. Conclusion                              1107
V. The Charge to the Jury                      1107
VI. Sufficiency of Evidence to Support the
Verdict                                     1110
VII. Admissibility of the FAA Airworthiness
Directives                                   1110

OPINION AND ORDER

EDWARD R. BECKER, District Judge.

I. Preliminary Statement

This is a suit on a policy of insurance covering accidental death. The jury, instructed as to the New York presumption against suicide which we held applicable to the facts, found for the plaintiff, Virginia Barry Melville, who was the sole beneficiary under the $500,000 policy. This opinion addresses defendant's post-trial motions.^[1] At the core of those motions and of the case lies a complex question of choice of law: whether the presumption to be applied with respect to suicide is that of Pennsylvania, New York or Delaware where: (1) Pennsylvania is the forum state, the situs of the principal office of the insurance broker which placed the policy, and the sometime residence of the plaintiff; (2) New York is the place where the insurance contract was made, as well as the state where the defendant insurer was incorporated and had its main office; and (3) Delaware was the residence of the insured, the situs of his death, the locale where most of the facts relevant to the question of accident or suicide occurred and the place where the order for insurance was initiated. In answering that question we will be obliged to determine the applicability of the landmark case of Griffith v. United Air Lines, 416 Pa. 1, 203 A.2d 796 (1964) in a contractual context. In the course of that determination we will be obliged to engage in a comprehensive survey and analysis of Pennsylvania choice of law principles and their underlying policy bases.

The foregoing is a synopsis of the facts underlying our choice of law decision. The facts which formed the basis for the decision on the merits are not so mundane; indeed, they are extraordinary. They relate to the life and times of the insured, Josiah Marvel ("Jay") Scott, member of a respected and wealthy Wilmington family, who bilked the family fortune of many millions, most of which he squandered on disastrous venture capital undertakings, and some of which he lavished upon the plaintiff, who was his secretary and his mistress. The saga ended — and this case began — when Scott, on the brink of exposure for his misdeeds, died in an exceedingly strange airplane crash which was presaged, in full view of spectators at the Wilmington airport, by bizarre aerobatic maneuvers. The trial testimony thus ranged from the moods and escapades of Jay Scott to the aerodynamics of aerobatic flight. The aerodynamic testimony was adduced to aid the jury in determining whether the "loops," "split-S," "tight turns" and "hammerhead stalls" performed by the chartered Atlantic Aviation Piper Cherokee Arrow in which Scott was a passenger at the time of the fatal crash occurred, as the plaintiff contends, either because the pilot was responding brilliantly to a mechanical malfunction, or because the pilot was showing off or had gone mad; or, as defendant argued to the jury, because Scott had interfered with the pilot's use of the plane's dual controls in order to cause the plane to crash, and the pilot had valiantly, but futilely, attempted to compensate for Scott's actions so as to keep the plane aloft.

The choice of law question, which is the principal issue considered in this opinion, is best understood in context of the procedural history of the case and of the facts adduced at trial; hence, we address these matters first. However, we shall also take up the following issues raised by defendant's motion:

(1) whether the evidence was insufficient to support the verdict, so that defendant was entitled to a directed verdict and, thus, is now entitled to a judgment n. o. v.;

(2) whether the verdict was so contrary to the weight of the evidence that defendant should be granted a new trial;

(3) whether we correctly instructed the jury as to the New York presumption against suicide; and

(4) whether we erred when we admitted into evidence certain F.A.A. documents (Airworthiness Directives) which opined that particular parts of the Piper Cherokee Arrow model involved posed dangers serious enough to require prompt and/or regular inspections, repairs, or replacements.

The choice of law question is both unsettled and abstruse. However, for the reasons which follow, we reaffirm the choice of law decision previously made, although on different grounds from those on which it was previously made. In the course of our discussion we will note and document our conclusions that:

(1) the presumption against suicide is substantive not procedural, hence not to be determined by mere reference to the law of the forum;

(2) in general terms, the post-Griffith Pennsylvania choice of law cases have moved from (Griffith) interest analysis to a hybrid scheme containing elements of both interest analysis and the "significant contacts" approach of Restatement (Second) of Conflict of Laws (Restatement II), but with strong territorial overtones reminiscent of the traditional rules of Restatement of Conflict of Laws (Restatement I);

(3) in contract choice of law cases specifically, the post-Griffith Pennsylvania case law contains a strong Restatement I strain, i. e. numerous cases have retained the traditional territorial (place of contracting) rules, though the traditional approach vies with the post-Griffith interest analysis/Restatement II hybrid for superiority;

(4) in any event, the post-Griffith case law on choice of law in contract cases is in utter disarray and confusion, leaving unclear how Pennsylvania would resolve the case at bar, particularly in view of the fact that, in interest analysis terms, this is an "unprovided-for case," i. e. one in which no state has an interest;

(5) it is therefore necessary to examine the issue before us on alternative assumptions: that the traditional choice of law rules for contracts are applicable, and that the Griffith interest analysis/Restatement II hybrid approach is applicable; and

(6) both assumptions lead to choosing New York's law on the presumption of suicide.^[2]

We also are satisfied that we correctly stated New York law when we charged the jury that if they believed there to be a "fair question" whether Scott died from accident or suicide, they should "find accident." Additionally, we find no error in our evidentiary ruling(s) and hold the evidence to be sufficient to sustain the verdict.^[3] Accordingly, defendant's post-trial motions will be denied.

II. The Procedural History

The litigation between the parties did not begin in this Court. Instead, plaintiff first sued the defendant in the state courts of New York. That suit was dismissed on the ground of forum non conveniens, at least in part because the insurer claimed that most of the witnesses who could testify as to whether the crash was an accident or the product of a suicide resided in Delaware or Pennsylvania. The dismissal was conditioned upon the insurer's consenting to jurisdiction either in Pennsylvania or in Delaware and on its waiving the statute of limitations. Barry v. American Home Assurance Co., 38 A.D.2d 928, 329 N.Y.S.2d 911; aff'd 31 N.Y.2d 684, 337 N.Y.S.2d 259, 289 N.E.2d 180 (1972). Thence, Ms. Melville, who was then a citizen of Pennsylvania, brought this diversity action.

III. The Facts Adduced at Trial

The facts adduced at trial concerned three subjects: (1) the placement of the insurance policy; (2) the events in Jay Scott's life preceding the fatal flight; and (3) the flight itself, (including both eyewitnesses and expert testimony).

The evidence with respect to the insurance policy was undisputed and was essentially as follows. In 1966, Scott, a citizen of Delaware, ordered the insurance policy through the Delaware office of Johnson & Higgins, an insurance broker whose main office is in Philadelphia. As Scott had requested, the broker placed the order by phone call with the defendant, a New York corporation, which effected an oral binder at defendant's principal place of business in New York. Defendant subsequently issued the written policy and posted it in New York, mailing the policy to the broker's Philadelphia office, through which it eventually reached Scott in Delaware. That policy was in effect at the time of Scott's death and is the subject of this case.^[4]

Turning to the remaining facts, we shall attempt to present a balanced picture of what the jury heard. This will enable us to evaluate defendant's arguments that the evidence was insufficient as a matter of law and that the jury's verdict was against the weight of the evidence. In so doing, it is best to begin with the events of October 16, 1970, the fateful day on which an airplane crash took the life of Jay Scott and of Rudolph Halamka, the pilot of the chartered plane.

Halamka was an experienced, former navy pilot. Although his occupation was selling insurance, his true love was flying. He took charter assignments for Delaware Aviation Service whenever they offered him a flight. On October 16, 1970, Delaware Aviation called on Halamka to fly a charter from Wilmington, Delaware, to Teterboro, New Jersey (which is near New York City), and back again. He agreed. The only passenger in the single-engine plane was the charterer, Jay Scott.

The flight to Teterboro was uneventful. Scott disembarked, went to New York where he lunched with a business associate, Quing Yng-Wong, and returned to Teterboro. Prior to boarding the plane, he placed a telephone call to the plaintiff Virginia Barry (Melville) to request that she meet him for dinner upon his return to Wilmington.^[5] Ms. Barry had been Scott's secretary for a number of years. She was, as we have noted, also his mistress, and he had kept her in style for quite some time.^[6]

On the return flight to Wilmington, after it had been cleared for landing, and while it was flying at an altitude of less than 1000 feet over a congested area near the airport, the plane engaged in a series of what appeared to be aerobatic maneuvers: one or two loops, some steep turning climbs and dives, a split-S, and one or two hammerhead stalls.^[7] The plane failed to recover from a hammerhead stall and crashed nose first into the ground, killing both Halamka and Scott.

The eyewitnesses disagreed to some extent in their characterization of the maneuvers. Plaintiff called Mr. Kirk, the Wilmington Airport air traffic controller, who testified that the flight appeared to be "coordinated;" i. e., it appeared that the three basic control surfaces of the airplane were being operated as a unit.^[8] However, defendant presented an eyewitness, Mr. Meihofer, an experienced pilot, who insisted that the flight was erratic and uncoordinated and that he had also seen the plane "porpoising."^[9]

There was no dispute that the maneuvers, whether "coordinated" or not, were extraordinarily dangerous. The plane, a Piper Cherokee Arrow, was not stressed for aerobatics, and its altitude was too low to provide the needed safety margin for aerobatic flying. Moreover, F.A.A. regulations forbid performance of aerobatics in a plane which is either within the control area of an airport or in a congested area. The Arrow was flying in a locale which rendered it in violation of both proscriptions. Plaintiff made no attempt to dispute defendant's claim that if a pilot were caught flying in this manner, he would lose his license, unless he had an extraordinary excuse for same.

So then, there was no dispute that the observed flight was bizarre and hazardous. Indeed, the Wilmington air traffic controller testified that he had not seen such flight in a congested control zone, at such low altitude, in his 12 years on the job.^[10] Another expert insisted that he had seen such maneuvers done as "show-boating" on several occasions; but he did not suggest that he had seen such a performance by a pilot who had been cleared for a landing, and who was therefore identifiable and sanctionable.^[11]

Additionally, defendant presented substantial evidence that Halamka was a competent pilot, stable and happy, with no known reason to risk his license, his reputation, or his life. In this regard, defendants called Mrs. Halamka, the pilot's widow. Her testimony, and that of others acquainted with Halamka's character and skill, revealed no flaw that would suggest his responsibility for the crash.^[12]

Both plaintiff and defendant offered expert aeronautical testimony to explain the crash. We shall summarize that testimony in due course. The defendant did not, however, rely solely, or even principally, upon that testimony, except to forge a link in its theory that Jay Scott, then beyond desperation, took his own life (and that of the pilot) to avoid the shame of his gross misdeeds against his family and social peers, the imminent total exposure of which would cause his world, already in shambles, to crumble. The aeronautical testimony merely closed the ring by showing how, according to defendant's theory, Scott was able to commit suicide, while making it look like an accident. In view of the predominance of the social history in defendant's version of the case, we begin by summarizing the extensive testimony which the defendant offered to establish it.

After a stint in the Marines, Jay Scott received his B.A. from Yale and made a start towards a career in law by entering the University of Virginia Law School. However, by Christmas his law career had ended. Jay so disliked law school that he left and began training as an investment counsellor, under the tutelage of a close friend of his parents who managed the family finances. Over the years, Jay's parents had placed most of their multi-million dollar estate into a variety of trusts for themselves, and for their three children, Jay, his brother and sister.

By 1960, Jay was given full responsibility for financial management of the family trusts. By 1970, he had misappropriated several million dollars from the trusts and apparently dissipated every penny he had taken. He obtained the trust corpus by forging the signatures of his parents to stock assignments and by causing his secretary, the plaintiff, to do likewise. Cleverly, he concealed his misappropriations by regular and timely payments of "dividends" to the family calculated to match the dividends of the stocks that were supposedly the corpus of the trust. Jay invested the proceeds of his sales of the family fortune in high risk "venture" stocks, as well as in high living. By 1970, almost all the investments were in companies that were going bankrupt. Meanwhile, Jay had also fraudulently obtained substantial loans from several Wilmington banks and had engaged in an extensive pattern of kiting, writing checks on insufficient funds. He had even used his position of trust as Treasurer of the University of Delaware Library Association to misappropriate some $34,000. The defendant adduced evidence at trial that at the time of his death Jay Scott defendant owed a total of three million five hundred thousand dollars ($3,500,000) to 27 creditors. Among the largest creditors were: (1) Capital Bank and Trust Company . . . $521,147.17; (2) Chandler Leasing Division, Pepsico Leasing Corporation . . . $440,040.82; and (3) Delaware Trust Company . . . $327,271.44.

Only 8 days before the fatal crash, Jay's underground life began to surface. The Scott family attorney, learning through his firm's representation of a Wilmington bank that large judgments had been entered against Jay, informed Jay's uncle. Together they told Jay's parents, who in turn told Jay's brother. The parents and brother were, of course, concerned and sought an explanation. Jay's mother, dissatisfied with his explanation, decided to reassure herself by checking the safe deposit box in which the family trust securities were kept. It was empty.

When Mrs. Scott confronted him with this news, Jay lied, telling her that all the securities were in her personal safe deposit box. Mrs. Scott, not having used the box in years, could not locate her key, and Jay conveniently claimed to have lost his. When the bank drilled Mrs. Scott's box open, only three days before the crash, it was also empty. When he was confronted with this news, Jay lied again. Knowing that his father was away on a trip, he told his mother that all the securities were safely in his father's box. By this time, in defendant's view, Jay was trapped, having burned all his bridges behind him. Defendant theorized further that Jay thereupon laid plans for a suicide-murder that would prevent his having to face the disgrace which would ineluctably follow from the discovery that he had stolen millions from his mother, father, sister, and brother, and that he had defrauded a charitable endeavor in which he was associated with other prominent Delawareans. Moreover, he, and perhaps Virginia, the plaintiff, faced criminal prosecution. And, while much of the over $2 million in life and/or accidental death insurance which Jay possessed had been purchased some years before and not in the face of this crisis, defendant theorizes further that, because of the insurance, Jay could plan that his wife and his mistress and his children by both of them would be provided for.

Defendant's theory that Jay planned an airplane suicide was buttressed by the following evidence about his final days. First, one week before the crash, on October 9, 1970, Jay took out a $50,000 insurance policy covering only the risk that he die accidentally while travelling and naming a new mistress as beneficiary. Moreover, Jay acted suspiciously on a charter flight which he took on Monday, October 12, 1970, one day before his mother was scheduled to have her (empty) safe deposit box opened and only four days before the crash. As on the final day, Jay had chartered a flight to Teterboro, but due to weather conditions, he was carried in a twin-engine two pilot plane, in which the passenger sat in the rear seat. Jay proceeded to ask questions about the controls of the plane and about how they compared to the controls of a single-engine plane. After Jay disembarked from the plane, the two pilots saw him examining the control panel of a single-engine plane.^[13] The next day Jay ordered the fateful charter, insisting that a single pilot single-engine plane be provided.^[14]

Admittedly, Jay left no suicide note. In fact, all the witnesses except one, both plaintiff's and defendant's agreed that he showed no signs of despondency or of impending doom in the weeks, days and hours preceding his death. Even at the luncheon which immediately preceded the final flight, Quing Yng-Wong reported that he and Scott had kidded around.^[15]

Defendant, however, says that this is not evidence contra suicide, but rather evidence of how coldly Jay Scott planned. Indeed, defendant submits, Scott's activities and personal contacts in his last day were so preposterous, given the imminence of his exposure, that they can only be understood as a charade, a deliberate, if overdrawn, attempt to leave evidence that he had not committed suicide so that his beneficiaries could collect on his insurance policies. Thus does defendant account for Jay's call to the plaintiff before he left Teterboro airport to return to Wilmington, and the substance of his luncheon earlier that day with Quing Yng-Wong at which Jay evinced a serious continuing interest in his doomed business ventures and attempted to get Quing Yng-Wong to aid therein. More representative of Jay's true state of mind, defendant contends, is his encounter a few weeks prior to the fatal crash with a personal friend who was then the Mayor of Wilmington. During that encounter Jay, looking defeated, confided that he was in "deep trouble," and importuned the friend for a $75,000 loan, which was promptly made.^[16]

As we suggested above, the significance of the defendant's interpretation of the "motive" evidence depends on its presenting a plausible theory of how Scott could have caused the plane to crash that matches the eyewitness accounts of how the plane did in fact crash. Defendant's theory is simple. It is that Scott, having been told only a few days before how the various controls of an airplane work, waited until the plane was near to landing and then used the controls to crash it.

According to defendant's theory, the reason that Jay had insisted on a single engine plane was that in such a plane the pilot and the passenger sit behind the dual set of controls. Thus the passenger could push or pull on the control yoke causing the plane to climb or dive; he could turn the yoke right or left causing the plane to bank; and/or he could step on the left or right rudder causing the plane to move left or right. Halamka, a competent pilot, would have been applying just the right pressure on any or all of these controls to direct the plane in for a landing. A sudden strong pressure, exerted by Jay Scott, in either direction on any of these controls, would have led to a sudden change in flight. Halamka could have compensated by struggling to use the same or a different control, attempting to overcome Scott's action. Defendant concludes that the result of such a struggle would have been precisely what the eyewitnesses saw: aerobatic maneuvers with some degree of coordination, but done sharply and erratically. Thus theory as to how Jay actually accomplished his suicide-murder was supported by the testimony of Mr. Meihofer, an eyewitness to the crash who is himself a pilot, and of Mr. Robinson, called by defendant as an expert in aeronautical engineering.

Counsel for the plaintiff, Ms. Melville, presented a different version of the events at issue. He conceded the defendant's chronicle of Jay's bad deeds, raising only a faint defense to the near certainty that Jay had been guilty of forgery,^[17] and none at all to his possible prosecution for embezzlement. Neither did he deny or seek to explain away Jay's desperate financial situation. Plaintiff's theory, to counter defendant's strong evidence, was severalfold.

First, plaintiff advanced the view that Jay Scott, the intrepid one, possessed a personality which would have enabled him to weather these storms. Second, with the exception of his conversation with the friend who had made the $75,000 loan, plaintiff argued, with little contradiction, that Jay's frame of mind appeared sanguine, and that he was looking to the future. Third, plaintiff relied on the law's presumption against suicide, which, deeply rooted in human experience, rests on the notion that man, instinctively or consciously, clings to life.^[18] The presumption applies with magnified force in this case, plaintiff says, because if Jay Scott committed suicide, he also committed murder (of the pilot, Halamka). In moral, if not in legal terms, plaintiff argued most forcefully that there must be some kind of presumption against murder, too. And fourth, plaintiff relied upon expert testimony about the aerodynamics of aerobatic flight which may be summarized as follows.

During plaintiff's case-in-chief, George H. Rhodes, a pilot, flight instructor, and producer of radio and T.V. shows about flying, testified in some detail about the principles of flight and the effects of the use of various aircraft controls. Principally, he described the complex combination of pilot actions necessary to produce acrobatics — even those that appear to be simple. On the basis of this complexity and of his experience teaching new pilots, Mr. Rhodes expressed the firm opinion that a person who had never piloted a plane could not have performed the reported maneuvers.

Later, during rebuttal, plaintiff called a second expert, Wilbert N. Hubin, professor of physics, pilot, and teacher of aerobatics. Dr. Hubin testified in detail about the complexity of the acrobatic performance that preceded the crash. He agreed with Mr. Rhodes that a non-pilot could not have performed the reported maneuvers. Additionally, he rejected as impossible defendant's theory, developed through the testimony of Mr. Meihofer, that the aerobatics were the result of Halamka struggling to keep control of the airplane while Scott attempted to crash it. In fact, Dr. Hubin testified that if Scott had determined to commit suicide by use of the airplane's controls and had set about learning even the most rudimentary aspects of flight control, he would have realized that all he needed to do was to push the yoke forward causing a steep dive and a crash within 15 to 25 seconds, too little time for a pilot to react and save the plane.^[19]

Finally, Professor Hubin presented an alternative theory of the crash, namely, that most, but not all, of the maneuvers testified to could be explained by a mechanical malfunction in the "stabilitator" and by the pilot's attempt to compensate therefor. He suggested that the stabilitator, which aids in the control of pitch on the Arrow, could have become stuck in one position with the result that the plane's nose would keep coming up. If a skilled pilot were to find himself suddenly facing such a problem, there would be several alternative maneuvers he could perform to try to maintain flight while solving the problem. Dr. Hubin testified that the observed maneuvers would have been precisely the proper ones to execute. Dr. Hubin conceded that he could not prove to a certainty that such a mechanical malfunction had occurred; however, he insisted that his theory was not mere speculation. In support of his theory he cited two F.A.A. Airworthiness Directives, which reported that there had been mechanical failures related to the stabilitator system of the Arrow.^[20] Additionally, Dr. Hubin claimed that Exhibit D-38, a photograph of the wreckage, revealed that a bolt was missing; that the crash would not have caused that bolt to fall out; and that that defect could have led to the hypothesized mechanical failure.^[21]

Defendant countered Hubin's theories on rejoinder. At that time it called Arnold Robinson, an experienced aeronautical engineer, who disputed Professor Hubin's contentions by testifying that (a) the observed maneuvers were not consistent with and could not possibly have resulted from the suggested mechanical malfunctions, and (b) that the maneuvers were consistent with the hypothesis that the pilot was seeking to prevent a crash which the passenger was seeking to cause.

The record is voluminous; the trial lasted 9 days. While the foregoing survey of the evidence is incomplete, we believe it sufficient for present purposes, given the strictures of a judicial opinion. We will in due course (p. 1110 infra) address the question whether this evidence is sufficient to support the plaintiff's verdict. However, that task is best performed after discussion of the propriety of the jury charge (against which sufficiency must perforce be measured). And, since the correctness of the charge is in turn very much a function of which state's law governs, we must turn now to the choice of law questions.

IV. Choice of Law

A. Introduction

The choice of law issue has dominated this litigation since its inception. Indeed, it seems likely that both plaintiff's initial choice of a New York forum and defendant's forum non conveniens motion which subverted that choice were motivated, at least in part, by opposing counsels' jointly held belief that New York law places insurers at a great disadvantage in suits on accidental death policies.^[22]

The question before us is whether, as a federal diversity court applying the forum state's substantive law, Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938), including its choice of law principles, Klaxon Co. v. Stentor Electric Manufacturing Co., Inc., 313 U.S. 487, 61 S.Ct. 1020, 85 L.Ed. 1477 (1941), we should apply the law of New York, or of Delaware, or of Pennsylvania in dealing with the presumption against suicide in this case. The presumption against suicide is, as we have noted, the judicial expression of the venerable and deeply rooted human judgment that man instinctively as well as consciously clings to life and that, notwithstanding the discoveries of modern psychiatry, suicide is contrary to the primordial reflexive natural will to live.^[23]

Plaintiff has championed New York law, construing it to place the risk of non-persuasion on the suicide issue in suits on accidental death policies on the insurer. Defendant, while not conceding that New York law goes that far, suggests that either Pennsylvania or Delaware law is applicable and that the law of these jurisdictions is much clearer in putting the burden of proving "accident" on the claimant. Knowing the importance of the choice of law issue to the conduct of the trial, we informed counsel that we would determine the law to be applied in advance of trial. After the parties had fully briefed their contentions, we ruled in a pre-trial bench opinion that we would apply New York law to the accident/suicide congeries of issues. Our rulings at trial and our jury charge reflected that decision. In responding to defendant's post-trial motions, we re-affirm that choice. We do so, however, on a different legal ground than we did previously.

The choice of law issues are extremely difficult. They require us to traverse much obscure terrain. They thrust us into heavy seas which are sometimes unmapped and at others charted with conflicting signals or directions. The problem begins with the Pennsylvania caselaw. The opinions of the Pennsylvania courts — both state and federal — have left Pennsylvania's choice of law rules and methodology with respect to contract cases in utter disarray; indeed, the courts have used facially inconsistent legal standards without acknowledging apparently conflicting precedent. The parties in their briefs have attempted to impose their own order on these cases; but their efforts have wrought neither synthesis nor consensus.

We now commence our analysis of the choice of law problem. The road to the ultimate conclusion is a long one, which will require us to touch many bases and to engage in a detailed analysis of the policy roots of potentially applicable choice of law rules. It will also involve us in extensive discussion of the "vested rights" methodology of Restatement Conflict of Laws (Restatement I), the "most significant relationship" approach of Restatement (Second), Conflict of Laws (Restatement II), and the "interest analysis" of Griffith v. United Air Lines, Inc., 416 Pa. 1, 203 A.2d 796 (1964). Our choice of law discussion will proceed as follows.

First, we will address the question whether the presumption against suicide as it is involved here is substantive or procedural. We consider this question first because, if the presumption is purely procedural, we may be obliged to apply Pennsylvania law as the law of the forum. See Sloniger v. Enterline, 400 Pa. 457, 162 A.2d 397 (1963); Carroll v. Godding, 155 Pa.Super. 490, 38 A.2d 720 (1944). But see discussion infra. Because, however, we find the presumption to be substantive we must face the difficult question to which the bulk of this opinion is devoted, i. e. which state's substantive law should be applied (as to the presumption of suicide). Second, and in pursuit of the answer to that question, we shall consider whether there are in fact such conflicts among Pennsylvania, New York and Delaware law on the issues of burden of proof and/or presumption against suicide in a suit to collect on an accidental death insurance policy to make choice of law a determinative issue. We shall find that there are. Third, we will briefly outline the salient points of the three competing choice of law systems which have influenced Pennsylvania choice of law jurisprudence: Restatement I, Restatement II and "interest analysis." This outline is a necessary background to our subsequent discussion. Fourth, we shall review the conflicting lines of cases on Pennsylvania choice of law both in the Pennsylvania appellate courts and in the United States Court of Appeals for the Third Circuit which bear upon the application of Griffith's interest analysis to contract cases. (The discussion shall perforce not be limited to contract cases). In the course of this discussion we shall explain the disarray and confusion in the law to which we have referred, as well as draw our own inferences as to what choice of law methodologies a Pennsylvania court in 1977 would apply to the problem before us. Fifth, we shall rescribe the parties' choice of law contentions; Sixth, we shall explicate the basis for (and the defects in) our prior bench opinion; and Seventh, we shall apply our inferences as to Pennsylvania choice of law methodology to the instant case, reaching the same choice of law result as we did at trial, namely, that New York law applies.

B. Is the Presumption Against Suicide Procedural or Substantive?

We commence our discussion with reference to the traditional choice of law rules. Under these rules questions of burden of proof and of the existence and the significance of presumptions are generally classified as matters of procedure, Restatement I, § 595, and matters of procedure are governed by the law of the forum, id., § 585. The presumption against suicide, including the New York "fair question" formulation, is susceptible to being characterized either as a matter of interpretation and construction or as a matter of procedure, burden of proof, and presumption. Although the Pennsylvania choice of law cases have necessarily engaged in classification,^[24] they have rarely conceded aloud the importance and difficulty of that function. For a rare example, see Griffith (tort or contract).

There have been myriad decisions in which the Pennsylvania courts have classified presumptions or burden of proof as procedural rules, and have applied lex fori. See Sloniger v. Enterline, 400 Pa. 457, 162 A.2d 397 (1960). See also, Rodney v. Staman, 371 Pa. 1, 89 A.2d. 313, 315-316 (1952); Carroll v. Godding, 155 Pa.Super. 490, 38 A.2d 720 (1944). Yet we do not believe these cases compel us to apply lex fori here because they do not hold that where a facially procedural rule has substantive as well as procedural aims lex fori must govern. Indeed, in Sanders v. Glenshaw Glass Co., 204 F.2d 436 (3d Cir.) cert. denied 346 U.S. 916, 74 S.Ct. 278, 98 L.Ed. 411 (1953), the Third Circuit found that Pennsylvania choice of law referred to Massachusetts tort law and thereby adopted the Massachusetts burden of proof. It may have been that the Court did so because it was so closely tied to the emerging substantive liability of a products manufacturer. At any rate, neither Sloniger nor Sanders offers us any clear guidance to the classification problem.

As we will explain in detail below (see note 66 infra) the fact alone that a state is the forum is an inadequate ground for asserting legislative jurisdiction^[25] over events outside the forum. Yet not even the most stringent (Restatement I) territorialist would recommend either that a forum must adopt all the procedural rules of the place where the disputed events occurred or that all causes of action be strictly local, suable on only in one jurisdiction. It is simpler to allow that the forum's basic structure of decision-making be left generally untouched, while the substantive law of another jurisdiction is adopted. However, as the progeny of Erie have proved, see, e. g., Hanna v. Plumer, 380 U.S. 460, 85 S.Ct. 1136, 14 L.Ed.2d 8 (1965), Ragan v. Merchants Transfer & Warehouse Co., 337 U.S. 530, 69 S.Ct. 1233, 93 L.Ed. 1520 (1949), Guaranty Trust v. York, 326 U.S. 99, 65 S.Ct. 1464, 89 L.Ed. 2079 (1945), there is no simple litmus test of procedure and substance. We turn, then, for our solution to the classification problem, as we believe Pennsylvania would do, to the collective wisdom of those courts and commentators which have considered the substance-procedure classification problem in choice of law in contexts akin to the case at bar.

In Pilot Life Insurance Co. v. Boone, 236 F.2d 457, 462 (5th Cir. 1956), Judge Rives spoke for the panel in reversing a district judge's ruling that the forum (Alabama) would apply its own presumption against suicide in a suit on life insurance policy, although it would defer to lex loci contractus (South Carolina) on "interpretation."^[26] Acknowledging that presumptions are ordinarily classified as procedural and decided under lex fori, Judge Rives classified this "presumption" as a matter of substantive contract law. He found that "the effect of this presumption against suicide is . . . inseparably connected with the substantive right to defend . . .." 236 F.2d at 462-463. Judge Rives reasoned that

[W]hile many cases can be found to the effect that ordinary presumptions are procedural and governed by the lex fori, we think that it would be a misnomer to bring this presumption against suicide within that class.

In Buhler v. Maddison, 109 Utah 267, 176 P.2d 118, 168 A.L.R. 177 (1947), the Utah Supreme Court conceded that courts were divided on the issue whether rebuttable presumptions and burden of proof were substantive or procedural but concluded that the trend was to classify as substantive any matter on which "there may exist a difference of opinion," quoting from Annotation, 78 A.L.R. 883. We believe this to be a salutary approach. But see Weber v. Continental Casualty Co., 379 F.2d 729 (10th Cir. 1967) holding that Oklahoma would not apply the (arguably substantive) presumption of a locus state.

In addition to the evidence of cases such as Pilot Life and Buhler, we note that both Restatement II and the Federal Rules of Evidence have reached similar results. Thus Restatement II authorizes use of forum law as to the burden of proof and the effect of presumptions "unless the primary purpose of the relevant rule of the state of the otherwise applicable law is to affect the decision of the issue. . . ." §§ 133 and 134.^[27] The presumption against suicide appears to us to be generally designed to "affect the decision of the issue." The case at bar illustrates that point dramatically. Moreover, the presumption against suicide appears to us to "single out a relatively narrow issue from the general norm." See Comment to Restatement II recited in note 26 supra. Fed.R.Evid. 302 is to the same effect in requiring that

the effect of a presumption respecting a fact which is an element of a claim or defense as to which state law supplies the rule of decision is determined in accordance with state law.

(This means that state substantive law, under Erie, rather than federal procedural law supplies the rule). The presumption against suicide seems to us to respect a fact which supplies an element of a claim or defense.^[28] Moreover, we find additional support in McCormick's view that burdens of proof and presumptions "are almost always allocated for the primary purpose of affecting the decision in the case . . . where the jury is in doubt. To say that these rules merely govern the conduct of the trial . . . gives far too much emphasis to form over substance." McCormick, Evidence 835 (2d Ed.1972).

We agree with Pilot Life, finding it difficult to conceive of a "presumption" more substantive in nature in terms of affecting the substantive claim or defense than the presumption against suicide. And what Judge Rives found to be true of the presumption against suicide generally we believe to be especially true in this case. For instance, after analyzing New York's law, we will find that the New York Court of Appeals now regards the presumption against suicide as a "rule of jurisprudence . . . which describes a basic value judgment of our system," People v. Levya, 38 N.Y.2d 160, 169, 379 N.Y.S.2d 30, 37, 341 N.E.2d 546, 551 note 3 (1975), which may mean it is a conclusion of substantive law employed when the facts of suicide or accident are in doubt, and not a "presumption" at all. See pp. 1082, 1107-1110 infra. Therefore, New York's "presumption" may be as "substantive" as any other common law in New York. If this interpretation of the New York presumption is correct, the mere fact that Pennsylvania, whose choice of law principles we are following, does not have such a counterpart substantive rule, p. 1081 infra, does not preclude Pennsylvania from classifying the New York presumption as "substantive" for choice of law purposes. Quite the contrary, the situation would be identical to what occurs in choice of law every time a forum state has no substantive law on the issue and a sister state does have one: a "conflict" is presented and the forum must decide, under its choice of law principles, whether to apply that sister state's substantive law or not.

As the Pennsylvania Supreme Court in Griffith refused to be blinded by "[m]ere technicalities of pleading," 416 Pa. at 11, 203 A.2d at 800, so we believe that Court would refuse to be bound by the label "presumption" and would recognize that its legal difference with New York is a matter of substance not of procedure.^[29] We are supported in this view by Pilot Life, Buhler, and Weber, Restatement II, Fed.R.Evid. § 302 and McCormick. We need not go so far as to predict that Pennsylvania would actually adapt Fed.R.Evid. 302 to interstate choice of law in order to conclude, as we do, that Pennsylvania would apply New York's "fair question" doctrine to a claim under a New York insurance contract in which Pennsylvania was disinterested (see discussion p. 1101 infra), because Pennsylvania would recognize that the New York rule is one meant to create or to affect substantive rights of insurers and their insureds.

In view of our conclusion that Pennsylvania would treat the presumption as substantive we must consider which state's presumption it would apply.

C. Must We Choose? — The Relevant Similarities and Differences Among the Laws of Pennsylvania, New York, and Delaware

The first question posed when choice of law appears to be at issue is whether a choice must really be made, i. e., whether different sovereigns whose laws may arguably be applied to the case have laws which conflict in relevant ways. See Cipolla v. Shaposka, 439 Pa. 563, 267 A.2d 854 (1970). If we find no conflict, then we need not choose, i. e., decide which state's law applies. This would be true by definition and regardless of what choice of law methodology is preferred.

We will first describe the issues as to which there is no conflict among New York, Pennsylvania, and Delaware law. All three jurisdictions recognize the distinction between "life insurance" and "accident/double indemnity" insurance contracts, the former covering the risk of death generally, the latter covering only the risk of accidental death. Similarly, all three jurisdictions recognize the effect of this distinction on the burdens of pleading, of going forward with the evidence, and ultimately of persuading the triers of fact. In a life insurance policy, if suicide is in issue it must be excepted by the policy and pled by the insurer. Hence the burden of proving suicide as an affirmative defense to a suit on a life insurance policy is always on the insurer. See, e. g., Watkins v. Prudential Insurance Company, 315 Pa. 497, 508, 173 A. 644, 649 (1934). In contrast, accident (or double indemnity) coverage requires the putative beneficiary to plead and prove accident as a condition precedent to the insurer's duty to pay. See, e. g., Adams v. Metropolitan Life Insurance Company, 136 Pa.Super. 454, 7 A.2d 544 (1939).

The insurance contract at issue in this suit described itself as an "accident" policy. It covered loss "resulting from injury," and defined injury as follows: "bodily injury caused by an accident occurring while this policy is in force as to the person whose injury is the basis of claim and resulting directly and independently of all other causes in loss covered by this policy." Part III(A) of the policy consisted of an "Accidental Death Benefit" applicable "where injury results in loss of life of the insured." Part VI contained an exception: "This policy does not cover any loss caused by or resulting from: (1) suicide . . ."^[30]

It seems clear that the policy was an accident policy, giving rise to a duty by plaintiff beneficiary to plead and prove accident.^[31] Plaintiff accepted and met her initial pleading burden: "Fifth, Josiah Marvel Scott died as a result of injuries caused directly by an accident and independently of all other causes."^[32] At the same time she may be said to have accepted the initial allocation of the burdens of producing evidence of accident and of persuading the jury of accident by a preponderance of the evidence.

All the above was reasonably clear, uncontested, and therefore not at issue in the trial. What was sharply at issue was the effect of the "presumption against suicide" on plaintiff's burdens. It is precisely at this point that Pennsylvania and New York law diverge dramatically, raising the choice of law problem to which the bulk of this opinion is devoted.

McCormick distinguishes among four types of presumptions based on their respective effects on the burdens of proof. First, and weakest, is the "standardized inference" which aids a party in satisfying its own burden of producing evidence but does not shift that burden to the adversary.^[33] Second is the type of presumption that shifts the burden of producing evidence (the so-called "Thayer" theory of presumptions).^[34] Third is the type that shifts both the burden of producing evidence and the persuasion burden to the adversary (the so-called "Morgan" theory of presumptions).^[35] Finally, the "conclusive" or "irrebutable" presumption precludes the adversary from rebuttal.^[36] For McCormick, neither the first nor the fourth are truly presumptions, the first because it has no greater effect than any other rational inference used in fact finding, and the fourth because it is in reality a conclusion of law.^[37] Only the second and third are "presumptions" in a meaningful sense.

Pennsylvania's common law presumption against suicide was analyzed in Watkins v. Prudential Insurance Company, 315 Pa. 497, 173 A. 644 (1934). The court described a valid "reasonable inference against death being self-inflicted," but found there was no "sufficiently broad factual base" to "warrant its projection to the height of a presumption of law." 173 A. at 648. "It is merely a permissible consideration of the nonprobability of death by suicide." Id. The court held its effect was too slight to shift the burden of producing evidence. Id. 173 A. at 649. Of course, if it could not effect the burden of producing evidence it could not alter the persuasion burden, which remained throughout on plaintiff, and the court so held. Id. In terms of McCormick's four categories, Watkins' presumption against suicide clearly belongs to the first — it is simply a permissible inference, with none of the evidentiary consequences of a true presumption. McCormick himself so evaluated Watkins. McCormick, supra, at 811 note 74.^[38]

In contrast, New York's presumption against suicide is based on "the truth drawn from general human experience, that death by suicide is an improbability, that most men cling to life." Wellisch v. John Hancock Mutual Life Insurance Co., 293 N.Y. 178, 56 N.E.2d 540 (1944). Wellisch described this particular presumption as a "rule or guide for the jury":

The ``presumption against suicide' means that when death by violence is shown and an inference must be drawn by the jury as to suicide or not, then the jury should in justice and good conscience draw the inference of accident, not suicide.

56 N.E.2d at 543. In terms of its effect, the court said only "[N]or is it the sort of ``presumption' that serves only to shift the burden of proof and disappears from the case as soon as evidence to the contrary is offered." Wellisch was subsequently interpreted in Begley v. Prudential Insurance Company of America, 1 N.Y.2d 530, 154 N.Y.S.2d 866, 136 N.E.2d 839 (1956) in the following manner:

When death has resulted from violence, the presumption against suicide does more than shift the burden of proof and upon having done so disappears from the case; it continues to the end of the case and if a fair question of fact is presented as to whether death was due to suicide or accident, then the jury should answer accident.

154 N.Y.S.2d at 868, 136 N.E.2d at 841. Neither Begley nor Wellisch is clear as to the precise effect of the presumption against suicide. At minimum, they must be read as requiring the shifting of the persuasion burden (McCormick's third type above); it is also possible to read them as mandating a "conclusive" presumption as long as there are facts to support it. In Part V infra we shall explain the interpretation we gave to the cases in charging the jury. For now it is sufficient to realize that New York's presumption is much stronger than Pennsylvania's.

Delaware has never spoken on the issue of the effect of the presumption against suicide in an accidental death insurance policy. That leaves us with two alternatives. First we could make our determination of what Delaware law would be, reasoning from that state's law in regard to analogous presumptions. Second, we could apply Pennsylvania's conflict of law rule that when a sister state's law is unknown or unclear it is presumed to be the same as Pennsylvania's law. In re Trust of Pennington, 421 Pa. 334, 219 A.2d 353, 356 (1966). That would, of course, make Delaware's and Pennsylvania's presumptions against suicide identical. On balance, the second course appears more faithful to our Erie/Klaxon command. However, our ultimate conclusion that Delaware law is not applicable under Pennsylvania choice of law methodology moots the question of which of the two above alternatives is preferable.

This discussion has revealed a sharp difference between Pennsylvania law on the one hand, and New York law on the other, reinforcing the conclusion we reached in the last section, IV(B), that choosing one or the other presumption could have a dramatic effect on the outcome of the trial. The Pennsylvania presumption would only have aided plaintiff in meeting her initial burden of producing evidence. This was accomplished in our trial without the aid of a presumption, and once accomplished the presumption would have had no further effect. In contrast, New York's presumption would have had far more effect, by treating the existence of a fair question of fact as mandating a legal conclusion of accident (shifting the persuasion burden in the process).^[39] These differences between New York and Pennsylvania law require us "to choose."

D. The Three Competing Choice of Law Systems

During the past fifteen years, Pennsylvania conflicts of law jurisprudence as it applies to contracts cases has been influenced by three competing theories. Because of the critical role they play in understanding this case and the case law, and because the cases have not generally articulated the points of similarity and difference among the three, it is necessary to begin with a brief explication of each theory.^[40]

The first is Restatement of Conflict of Laws (1934) (hereinafter Restatement I), the underlying assumption of which is contained in § I: "[n]o state can make a law which by its own force is operative in another state. . . ." The implications of this assumption are that a forum qua forum cannot apply its law to events occurring beyond its borders and that a locus state cannot require another state to provide a forum for vindication of rights created under its law. However, if a forum is willing to determine a foreign-based cause of action, this rule would require it to apply the appropriate foreign law. The traditional view is known either as "territorialism" or the "vested rights doctrine," and in practice leads to a plethora of "jurisdiction-selecting" rules. The traditional method requires that the one and only properly concerned jurisdiction be determined merely by locating the "place of the wrong" in a tort case, the "place of contracting" in a contracts case, the "place of decedent's domicile" for issues of estate administration, et alia.

Far removed from the assumptions behind Restatement I is the "interest analysis" theory expounded chiefly, but not exclusively, by the late Professor Brainerd Currie. In contrast to the mechanical jurisdiction-selecting rules, interest analysis begins with an exposition of the substantive law of the states which appear to have an important connection with the transaction at issue. That law is analyzed in terms of its purpose and of the interest of the sovereign which it is perceived to serve.^[41] Only then can one truly answer the key question whether the interests of sovereign state X would be served by applying its law to the facts.

Professor Currie identified four basic case patterns. The first occurs where forum state X has an "interest" in applying its law and state Y has none, or vice versa. Currie labelled this a "false conflict," and advocated choosing the law of the only concerned jurisdiction. The second pattern occurs where both states have interests; Currie described this as a "true conflict." The reverse image of a true conflict is the "unprovided-for" case, where neither state has an interest. Finally, the "disinterested third state" pattern occurs where a state which itself has no interest is adjudicating a case in which two or more states do have interests. In the last case, Currie argued for dismissal on forum non conveniens grounds. However in both the "true conflict" and "unprovided-for" cases, Currie argued that forum law, lex fori, should apply. This followed from his basic assumption that forum law was presumptively valid, and should only be "displaced" when another jurisdiction demonstrated more interest in applying its law than the forum itself could show. (Currie initially argued that another jurisdiction could only demonstrate "more" interest when the forum itself had no interest. Later he modified this position, advocating balancing comparative strengths of interests when both jurisdictions were interested. See note 44 infra). His prescription of applying forum law was based largely on convenience and ease of administration.

[i]f, however, there were a general presumption in favor of applying the law of the forum, and a disposition to apply foreign law only where some useful purpose would be served thereby, counsel and court could know in advance that there would be no point in ascertaining the foreign law. . . . A real economy might therefore result.

Currie, Selected Essays on the Conflict of Laws, 156 note 85 (1963).^[42]

The third theory is Restatement (Second) of Conflict of Laws (1971) (hereinafter Restatement II). Although the final version of Restatement II did not appear until six years ago, tentative drafts were circulated as early as 1953. Restatement II's basic instruction is to choose the law of the "state of the most significant relationship." Significant relationship is determined by identifying the "contacts" each respective jurisdiction has with the case at bar. Hence the theory is widely described as "counting contacts," "grouping contacts," or finding the "center of gravity," though the three phrases are not strictly accurate since Restatement II counsels one to look at the quality, not merely the quantity of contacts. Restatement II is suffused with many of the territorial assumptions of Restatement I, evidenced throughout its complex, three-part structure. First, there are a series of specific jurisdiction-selecting rules, analogous to those of Restatement I; for example, § 192 on the law to be chosen for a life insurance contract, and § 148 for the law to be chosen in a suit for fraud and misrepresentation. Second, a series of territorial "contacts" are enumerated for each general type of action, such as property, torts and contracts. For example, § 188 identifies contract "contacts" which include, inter alia, place of contracting and place of performance. Finally, § 6 contains choice of law "principles" for use in evaluating such contacts. Principle 6(f) points to "certainty, predictability and uniformity of result" which were and are the underpinnings of vested-rights, territorial thinking.^[43] Countering these strong territorial currents are §§ 6(b) and (c), which counsel evaluating contacts through the "relevant policies of the forum and other interested states," in a manner virtually indistinguishable from Professor Currie's interest analysis. Thus Restatement II may be said to have one foot in the camp of Restatement I and the other foot in the camp of interest analysis.

E. Pennsylvania Choice of Law Doctrine; the Legacy of Griffith and its Impact on Contract Cases

Under Klaxon, the crucial conflicts issues must be decided under the Pennsylvania approach to or rules of choice of law; however, as we have already suggested, the Pennsylvania cases are neither clear, nor consistent. We shall now summarize the major currents in the Pennsylvania state and federal appellate court decisions, concentrating on the legacy of Griffith and its impact on contract cases.

1. The Pennsylvania Supreme Court and Superior Court Decisions

Pennsylvania's approach to choice of law questions prior to October 14, 1964 may be illustrated by the decision of the Superior Court in Varas v. Crown Life Insurance Company, 204 Pa.Super. 176, 203 A.2d 505 (1964). Varas was a suit for the surrender value of a life insurance contract. The court relied on Restatement I for its resolution of choice of law questions, specifically § 355 on the "place of performance" and § 356 on an optional promise to perform a contract provision. Varas was illustrative of what Pennsylvania's approach to choice of law questions generally, and choice of law in contract cases specifically, had been for many decades.

The modern choice of law era in Pennsylvania began less than a month later, when the Pennsylvania Supreme Court decided Griffith v. United Air Lines, Inc., 416 Pa. 1, 203 A.2d 796 (1964). Griffith provided a major re-examination and re-evaluation of the theoretical bases of Pennsylvania choice of law. Plaintiff sought to convince the Pennsylvania courts to apply their own law by characterizing the case as one of "contracts" to which the traditional "place of contracting" rule would apply. However, the court, insisting that the case was really a "torts" matter, focused on the issue whether it should set aside the cognate "place of injury" rule for torts of Restatement I. Speaking through Justice Roberts, the court did so, boldly rejecting the ancient rule of lex loci delicti "in favor of a more flexible rule which permits analysis of the policies and interests underlying the particular issue before the court." 416 Pa. at 21, 203 A.2d at 805.

The plaintiff in Griffith was the estate of a decedent domiciled in Pennsylvania at time of death; defendant was a Delaware corporation with its principal place of business in Illinois. The aircrash that took decedent's life occurred in Colorado, which limited damage recovery, while Pennsylvania did not. Under these facts, the Supreme Court found that Pennsylvania had an interest in applying its law of unlimited recovery (for benefit of decedent's surviving dependents domiciled in Pennsylvania), whereas Colorado had no local interest to protect by limiting recovery, and that Pennsylvania law was the proper one to apply. It is noteworthy, however, that the court also found that Pennsylvania had an "interest" because "the relationship between decedent and United was entered into in Pennsylvania." 416 Pa. at 24, 203 A.2d at 807. Thus both the kinds of legislative interests considered under Professor Currie's interest analysis and §§ 6(b) and (c) of Restatement II were juxtaposed with the territorial contacts of other parts of Restatement II. Significantly, both Professor Currie and Restatement II (among many others) were cited in the Supreme Court's decision to discard the "wooden application of a few overly simple rules based on the outmoded ``vested rights theory'", 416 Pa. at 13, 203 A.2d at 801, but the court did not rely on or explicitly adopt either Currie's doctrine or Restatement II.

Griffith, while addressing itself to choice of law in tort suits, announced that "[w]e are at the beginning of the development of a workable, fair and flexible approach to choice of law which will become more certain as it is tested and further refined when applied to specific cases before our courts." 416 Pa. at 22, 203 A.2d at 806. The question for us is the extent to which Griffith has been subsequently extended to contract cases, sweeping aside Varas and the tradition it represents.

Two years after Griffith, Justice Roberts wrote for the Court in McSwain v. McSwain, 420 Pa. 86, 215 A.2d 677 (1966). See note 42, supra. Applying interest analysis to the question of interspousal tort immunity he found that Pennsylvania, where husband and wife were domiciled, had a valid interest in imposing its standard for immunity whereas Colorado, the place of the underlying tort being sued upon, did not. Thus, like Griffith, only one state was found to be interested, and that state's law was chosen to govern. Justice Roberts noted in passing that Restatement II would have reached the same result, but he did not rely on that Restatement. Rather, in language that suggested that Griffith was applicable to all choice of law questions, he declared that "what should be sought is an analysis of the extent to which one state rather than another has demonstrated, by reason of its policies and their connection and relevance to the matter in dispute, a priority of interest in the application of its rule of law." 420 Pa. at 94, 215 A.2d at 682.

Justice Roberts again wrote for the Court in Elston v. Industrial Lift Truck Co., 420 Pa. 97, 216 A.2d 318 (1966). In Elston a Pennsylvania resident was injured by a fork-lift truck accident in New Jersey, and sued on a negligence theory the Pennsylvania corporation that had supplied his employer with the fork-lift. The supplier sought to implead and claim contribution from plaintiff's employer, whose business residence is not stated but who had a plant in New Jersey, under New Jersey's Workman's Compensation law. The issue was whether New Jersey or Pennsylvania law should govern on a question of third-party contribution in the context of workman's compensation. Contribution was permissible in Pennsylvania but prohibited in New Jersey. Justice Roberts found that both Pennsylvania and New Jersey had legitimate interests at issue — Pennsylvania in providing contribution for the benefit of a local corporation, New Jersey in preventing contribution from an employer already liable under its compensation scheme. Unlike Griffith and McSwain, then, which were "false conflicts," Elston presented a "true conflict."

While Justice Roberts finally applied New Jersey law because he found that New Jersey's interest in its workman's compensation scheme was "paramount," the conceptual basis for this application is not entirely clear. For, as we read his opinion, it rests partly on Restatement II, 216 A.2d at 323, and partly on territorial concerns that the tortfeasor who delivered the truck in New Jersey "had elected to transact business in New Jersey and having exposed itself to a policy and program which fails to give recognition to contribution [he] must be governed by that determination." Id.^[44] Thus Elston, like Griffith, exemplifies a "hybrid" system of interest analysis and territorial contacts which we perceive as evolving in Pennsylvania tort choice of law cases in Griffith's wake.

Shortly after Elston, the Pennsylvania Superior Court had an opportunity to apply Griffith in a contract context but declined to do so. In Eastcoast Equipment Company v. Maryland Casualty Company, 207 Pa.Super. 383, 218 A.2d 91 (1966), a New Jersey corporation sued an insurance company whose state of incorporation and principal place of business are not stated. The policy was issued and delivered in Pennsylvania, and the events that would give rise to liability took place in New Jersey. The court concluded that Restatement I and specifically Varas controlled in the situation, and applied the law of Pennsylvania where the contract was formed, opining that Griffith was limited to torts, and finding "no indication" that it had affected the ancient "place of contracting" rule. McSwain and Elston, which had preceded Eastcoast and were sufficient to give at least some "indication," were ignored. The court admitted that, were it to apply a Restatement II test, Pennsylvania's contacts would be found "minimal," and implied that New Jersey would probably have been the state with the "most significant relationship." 207 Pa. Super. 383, 218 A.2d at 91 note 5.

Eastcoast Equipment was followed by Justice Robert's third post-Griffith decision, In Re Hunter, 421 Pa. 287, 218 A.2d 764 (1966), a review of a final decree of a Pennsylvania Orphan's Court. A child had been born in West Virginia, and his natural mother had signed relinquishment forms in that state. The child, his adoptive parents, and the natural mother had all subsequently returned to Pennsylvania where they were domiciled at time of suit. The Orphans Court decided that West Virginia law governed the validity of the relinquishment forms because they had been executed there. The Supreme Court reversed because Pennsylvania "by reason of the subject matter of the dispute and the status of the parties as domiciliaries of this Commonwealth, has an overriding and continuing interest in the resolution of the issues in contention." 421 Pa. at 290, 218 A.2d at 767. Apparently relying on Restatement II principles, Justice Roberts found "there is no choice of law problem." Id. The case is also consistent with an implicit interest analysis, based on a determination that there is a false conflict hence no "problem" in choosing, when one state has an overriding interest by virtue of its domiciliaries and the other state has no interest.^[45] West Virginia's non-interest in the construction of its relinquishment forms is assumed rather than examined. At least tangentially, Hunter undercut Restatement I's lex loci contractus by rejecting as irrelevant the trial court's conclusion that the "agreement" was binding under the law of West Virginia, the place of contracting. Thus Hunter strongly, if obliquely, supported the view that Griffith was applicable to Pennsylvania choice of law generally.^[46]

Two months after Hunter, the Superior Court again, as it had done in Varas and Eastcoast Equipment, upheld the validity of Restatement I for contract actions. Crawford v. Manhattan Life Insurance Co. of N.Y., 208 Pa.Super. 150, 221 A.2d 877 (1966). Crawford was a suit between the beneficiary, a West Virginia resident, and a New York-based insurance company. The policy had been mailed from Pennsylvania, where the insurance company had an office, but the insured (beneficiary's husband) had signed an amendment in West Virginia. The court concluded that the "last act necessary" to make the contract effective was the signing of the amendment, which occurred in West Virginia. Using the formalistic reasoning of Restatement I, the court held that since "delivery," that last act, occurred there, West Virginia law was applicable.

Having decided the question by means of Restatement I, Crawford offered an alternative analysis under Restatement II and interest analysis (which it treated as synonymous). The court reasoned that Pennsylvania had no "interest" in applying its own standard of misrepresentation (the underlying substantive contract issue) in a suit between a West Virginia beneficiary and a New York company because Pennsylvania's only "contact" was the residence of the insurance agent. Crawford is significant for recognizing that Restatement II and interest analysis might conceivably apply to life insurance contracts, but it does not faithfully apply these methodologies. While using interest analysis to determine that Pennsylvania itself was not a concerned jurisdiction, Crawford ignores the fact that two other jurisdictions — West Virginia and New York — were concerned, and that therefore a "true conflict" was presented between those two interested states. Under Currie's interest theory, a forum non conveniens dismissal would have been indicated. Or if Elston, decided five months earlier had been followed, interest analysis should have been employed first, and if that method discerned a true conflict then territorial factors should have been consulted for a solution. Elston further instructed not looking at any one narrow territorial act — such as the place of signing an amendment — but the overall balance of territorial concerns. Crawford does not pursue any of these approaches.

Crawford was followed by Justice Robert's fourth post-Griffith choice of law decision, Kuchinic v. McCrory, 422 Pa. 620, 222 A.2d 897 (1966). Four friends, all Pennsylvania residents, had flown to Miami, Florida, for a football game. En route home, their private plane crashed in Georgia, and the three passengers' estates sued the pilot's estate for wrongful death. Georgia law required a standard of "gross negligence" for recovery, while Pennsylvania required simple negligence. The Court, finding it wholly fortuitous that the plane crashed in Georgia, held that Pennsylvania was the only jurisdiction properly concerned with imposing its standard of care. Kuchinic is significant in that it was the first case to use Professor Currie's term "false conflict" to describe the respective state interests. However, in the same paragraph it spoke of state "contacts," finding that Pennsylvania had numerous territorial connections including the "place where the host-guest relationship was established, where it was intended to terminate, and as the domicile of all four of the aircraft's occupants." These contacts were thereupon said to make Pennsylvania the "state with the most significant interest in defining the legal consequences attaching to the relationship . . ." 422 Pa. at 624, 222 A.2d at 900. That phrase's formulation, and the methodology employed, make Kuchinic the clearest example of the post-Griffith "hybrid" system, whereby pure interest analysis has joined with Restatement II's grouping of contacts. For under Currie's interest analysis, a false conflict would be sufficiently shown by the fact that all parties lived in Pennsylvania. The incremental weight Justice Roberts gave to the "places" where the relationship began and was to terminate is implicitly a function of Restatement II's territoriality.

After this flurry of decision-making — six Supreme Court opinions in two years — four years elapsed until Cipolla v. Shaposka, 439 Pa. 563, 267 A.2d 854 (1970). There a Pennsylvania plaintiff, a motor vehicle passenger, sued a Delaware defendant, his host driver, because of an accident in Delaware. Under the stipulated facts, Pennsylvania law would have permitted recovery, whereas Delaware law because of its "guest" rule, would have precluded it. The court quickly realized that both jurisdictions had "interests" at stake, and using Professor Currie's terminology once again, described the case as a "true conflict." In resolving this conflict, the court, while disclaiming the mere "counting" of contacts as not the correct Restatement II methodology, proceeded to do just that. Pennsylvania and Delaware each had a relevant "contact" because of their respective plaintiff-protecting and defendant-protecting policies. But Delaware had another "contact" because defendant's car was housed there. This was relevant because, according to the court, this factor would have affected insurance premiums of Delaware residents. Thus Delaware's contacts outweighed Pennsylvania's. Moreover, "[i]nhabitants of a state should not be put in jeopardy of liability exceeding that created by their state's law just because a visitor from a state offering higher protection decides to visit there." 439 Pa. at 567, 267 A.2d at 856. The court described this as a "highly territorial approach," and than quoted a federal district court's view in 1949 that "departures from the territorial view of torts ought not to be lightly undertaken." Id. Delaware law was thus applied. Justice Roberts filed a strong dissent, arguing that since the interests and contacts of Pennsylvania and Delaware were evenly balanced (he found the majority's counting process specious), the "better law," (see note 40 supra), should be chosen, which he deemed Pennsylvania's. Justice Roberts argued that the heavy territorial emphasis went counter to recent Pennsylvania choice of law decisions.

A year later In Re Danz was decided. 444 Pa. 411, 283 A.2d 282 (1971). Danz involved the validity of a gift contract executed in Germany with respect to funds held in Pennsylvania banks. The court, in one sentence, decided that German rather than Pennsylvania law applied in this situation "because that is the place where the contract was made and was to be performed." However, unlike Varas, Eastcoast Equipment and Crawford, no Restatement I provisions were mentioned. Nor was Restatement II relied upon. Griffith was not even cited. The only authority was a 1914 case, State Bank of Chicago v. King, 244 Pa. 29, 90 A. 453 (1914). No attempt was made to describe the respective interests of Germany and Pennsylvania, or to examine and weigh the "contacts" each jurisdiction had.^[47]Danz is singularly lacking in facts which would permit an application of either interest analysis or Restatement II. No mention is made of where the donor-decedent died, though it appears to have been Germany, nor where he was domiciled at time of death, though this was probably Pennsylvania. Both facts would be crucial under interest analysis and also would have been important "contacts" under Restatement II. Similarly, no mention is made of whether donees were domiciliaries of Germany, Pennsylvania, or some other jurisdiction at the time of gift, nor of whether decedent had any beneficiaries or heirs at law domiciled in Pennsylvania who would benefit from an increased estate if the claimed gift to the donees were declared invalid. Without these facts, it is impossible to know whether either Pennsylvania or Germany had "interests" in this matter. Thus it is abundantly clear Danz is not an interest analysis case. By the same token the above omissions preclude a thorough-going Restatement II analysis. To be sure, "place of performance" and "place of contracting" are two contacts made relevant by § 188 of Restatement II, but these contacts may or may not have been sufficient to establish Germany as the "place with the most significant relationship." Unquestionably, the result in Danz is consistent with Restatement I, although under that theory "place of contracting" would have sufficed, and the mention of another factor (i. e. place of performance) was superfluous. Danz plainly rests on territorial assumptions, but exactly which theory it utilized is unclear.

Finally, in Gillan v. Gillan, 236 Pa.Super. 147, 345 A.2d 742 (1975), the Superior Court purported to expand considerably the reach of Griffith. Gillan announced that Griffith "overruled a substantial body of Pennsylvania case law and adopted the new Restatement's approach to conflicts problems." (emphasis supplied). 345 A.2d at 744. As this passage implies, Gillan considered Griffith's methodology applicable to all conflict of law subjects, not just torts; but it interpreted Griffith as being based solely on Restatement II. In Gillan, both parties, formerly husband and wife, resided in Pennsylvania at time of suit. The former husband had breached a separation agreement which by its terms survived but was not incorporated in a divorce decreed by a New York court. The issue was whether he could collaterally attack the agreement on the grounds of collusion. The Gillan court set out two sections of Restatement II as applicable to its choice of law determination: the general principles of § 6 and the contract provisions of § 188. Thus the court clearly felt that Griffith had made Restatement II applicable to contracts. In applying these sections, the court found New York to have the "greater interest" because "the state where the divorce is to be procured has a pre-eminent interest in the preservation or dissolution of a marriage," 345 A.2d at 745, which the rules against collusion were designed to protect. The court did not examine any other "contacts" made relevant by § 188, such as the current domicile of the parties. Thus the Gillan court in theory relied on Restatement II, but in application performed its own version of "interest analysis." Its interest analysis, of course, was at variance with Currie's interest analysis; under the latter theory, the fact of residence of both parties in Pennsylvania would have made that state the only concerned jurisdiction, and a "false conflict" would have been present.^[48] The Gillan court did consult Pennsylvania law for the limited purpose of determining whether public policy allowed enforcement of the separation agreement by Pennsylvania's courts, but this determination was made, not on the basis of Restatement II, interest analysis, or even Restatement I, but rather the "jurisdictional" holding of Silvestri v. Slatowski, 423 Pa. 498, 224 A.2d 212 (1966), which we have analyzed at note 56 infra.

We began our review of the case law to determine whether Pennsylvania as a result of Griffith, has abandoned its ancient rule of lex loci contractus, as represented by Varas. We conclude that there is no clear answer to that question. Justice Roberts' six opinions demonstrate a trend of expanding the range of interest analysis (and Restatement II) methodology beyond its pure tort-liability application in Griffith. For example, while McSwain involved interspousal immunity, and Elston involved contribution among joint tortfeasors, Hunter involved adoption. Similarly, the recent Superior Court case of Gillan applied the "modern" methodology to a separation agreement. On the other hand, while both Hunter and Gillan are more analogous to contract actions than to tort actions, neither purported to overrule lex loci contractus as Griffith had squarely overruled lex loci delicti. Moreover, in the three clear contract actions, Eastcoast Equipment, Crawford, and Danz, the Superior Court twice explicitly, and the Supreme Court once implicitly, reaffirmed the vitality of Restatement I. In sum, incompatible inferences can be drawn from the case law, i. e. that Restatement I still applies to contract cases, and that the hybrid interest/Restatement II analysis applies.

It is also important to describe at this juncture our conclusion as to the meaning of the second inference above: that Griffith, instead of spawning a pure "interest" line of cases, has bred a "hybrid" system blending interest analysis and territoriality. Interest analysis is initially employed by the courts, and if under that analysis a "false" conflict is demonstrated (only one jurisdiction interested) that jurisdiction's law is applied. Territorial contacts of Restatement II are used to buttress the result. This is the method of Griffith, McSwain, Hunter, and Kuchinic. Similarly, if the initial application of interest analysis demonstrates a "true" conflict (both jurisdictions interested), territorial contacts are used to resolve the choice. In contrast to Professor Currie, who would have applied lex fori in a true conflict situation, Pennsylvania has twice encountered true conflicts (Elston and Cipolla) and each time has deferred to the law of the sister state. Each time deference was for territorial reasons: in Elston because the seller of the fork-lift had gone to New Jersey to transact business, and in Cipolla because the guest had been in Delaware on a motor trip.

2. The Court of Appeals (Third Circuit) Decisions

Because of our uncertainty about the question whether lex loci contractus has been abandoned by the Pennsylvania courts, we sought the guidance of the United States Court of Appeals for the Third Circuit. In view of the confusion in the Pennsylvania cases and the fact that a federal court's role in this area is predictive only, it is not surprising that we found the Court of Appeals opinions to reflect, hence exacerbate rather than solve, the confusion. Speaking generally, we discovered that the following cases provided at least some support for the view that Griffith left untouched the Restatement I tradition for contract cases: Jamison v. Miracle Mile Rambler, Inc., 536 F.2d 560, 562 (3d Cir. 1976); William B. Tanner Co., Inc. v. WIOO, Inc., 528 F.2d 262 (3d Cir. 1975); Craftmark Homes, Inc. v. Nanticoke Construction Co., 526 F.2d 790, 792 n. 2 (3d Cir. 1975); Daburlos v. Commercial Insurance Co., 521 F.2d 18, 20 at n. 2 (3d Cir. 1975); Pittsburgh Bridge and Iron Works v. Liberty Mutual Insurance Co., 444 F.2d 1286, 1288 n. 2 (3d Cir. 1971); Boase v. Lee Rubber and Tire Corp., 437 F.2d 527, 529-530 (3d Cir. 1971); First Pennsylvania Banking and Trust Co. v. U. S. Life Insurance Co., 421 F.2d 959, 962 (3d Cir. 1969).

On the other hand, an equally impressive string of cases lends support to the view that Pennsylvania has adopted either interest analysis or Restatement II for contracts cases: Kademenos v. Equitable Life Assurance Society, 513 F.2d 1073 (3d Cir. 1975), Siata International U. S. A., Inc. v. Insurance Company of North America, 498 F.2d 817 (3d Cir. 1974); Travelers Insurance Co. v. Davis, 490 F.2d 536, 542-543 (3d Cir. 1974); Neville Chemical Co. v. Union Carbide Corp., 422 F.2d 1205, 1210-1211 (3d Cir.) cert. denied 400 U.S. 826, 91 S.Ct. 51, 27 L.Ed.2d 55 (1970); Slaughter v. Philadelphia National Bank, 417 F.2d 21, 26 n. 8 (3d Cir. 1969); Scott v. Eastern Air Lines, Inc., 399 F.2d 14 (3d Cir.) (en banc) cert. denied 393 U.S. 979, 89 S.Ct. 446, 21 L.Ed.2d 439 (1968); Mannke v. Benjamin Moore & Co., 375 F.2d 281, 283 (3d Cir. 1967).^[49]See also Exchange National Bank v. Insurance Company of North America, 341 F.2d 673, 675 (2d Cir. 1965) (applying Pennsylvania conflict law in a case transferred under 28 U.S.C. § 1404(a)).

When the Third Circuit cases are subjected to individual scrutiny, they may be brought into more unified focus than our rough classification suggests; for example, some of the dichotomy may result from the evolution in the Pennsylvania state court cases, discussed above, the result depending upon when the Court of Appeals wrote. Moreover, as we have already noted, faced with conflicting strains in Pennsylvania law, the task of predicting what the Pennsylvania Supreme Court would do was extremely difficult, and it is understandable that one group of judges might predict differently than another, given the "signs" from the most recent state case. Finally, in many of the Court of Appeals opinions, the choice of law issue was not critical; indeed, the discussion in most of the cited cases is quite brief. However, since the choice of law question is critical to this case, we cannot content ourselves with less than an exhaustive analysis. We note, of course, that our function is predictive too, albeit in a compounded sense since we must also take cognizance of the (predictive) views of our Court of Appeals. We shall commence the necessary analysis by first considering the circuit decisions (which, of course, bind us) in several groups, organized by the temporal order of the Pennsylvania precedents which must be regarded as the exclusive source of law on this issue under Erie.

Seven of the Pennsylvania cases discussed above, Griffith, McSwain, Elston, Eastcoast Equipment, Hunter, Crawford, and Kuchinic preceded the first group of circuit cases which we shall consider. Earlier, we concluded that in Eastcoast Equipment and Crawford the Superior Court had denied the relevance of Griffith to contract cases and that in Hunter the Supreme Court had obliquely disagreed. We now consider seriatim the Circuit's reaction to those cases.

Mannke, Scott, and Slaughter relied on Restatement II or on Griffith as establishing the contemporary choice of law methodology. Thus in Mannke the court found Pennsylvania to be the "center of gravity." 375 F.2d at 283. And in Slaughter the court demonstrated an implicit awareness of the "hybrid" system by announcing that "Pennsylvania's approach . . . is to analyze the policies and interests underlying the particular issue before the court and then apply the law of the state having the greatest interest and most direct contacts with the problem." 417 F.2d at 26 n. 8 (emphasis supplied). In Scott, the Third Circuit en banc considered a wrongful death and survival action which was grounded in contract, not tort. The court declared: "[a]s the contractual claim against Eastern is rooted in Pennsylvania Law, it is governed directly by . . . Griffith." 399 F.2d at 22. The court undertook an interest analysis, deciding the situation was a "false conflict," and citing McSwain and Kuchinic. The "place of injury" in Massachusetts was held to "not necessarily" create an interest in Massachusetts. Id. Neither Mannke, Slaughter, or Scott mentioned, or distinguished, Crawford or Eastcoast Equipment.

In First Pennsylvania a panel of the Third Circuit ignored Griffith in deciding that Pennsylvania would refer to the law of the place of contracting in insurance contract cases. However, it relied neither on Eastcoast Equipment nor on Crawford, but instead reverted to Franklin Life Insurance Co. v. Bieniek, 312 F.2d 365, 367-368 (3d Cir. 1962), which in turn emanated from New York Life Insurance Co. v. Levine, 138 F.2d 286, 288 (3d Cir. 1943), which had used an explicit Restatement I methodology.

In Neville, the Circuit ascribed to Pennsylvania the "interest" approach in contract cases. Again it cited Griffith, but none of the post-Griffith state cases; it cited Slaughter, but not First Pennsylvania. Finally, in its actual application of Griffith, the court did not in fact analyze any "interests," but rather located "the place of negotiation . . . the principal place of business," and similar territorial contacts to determine Pennsylvania to be the "center of gravity" of the contract in a manner consistent with Restatement II. 422 F.2d at 1211.

One year later, in Boase, Judge Aldisert presented a more extensive consideration of the Pennsylvania cases, though he did not cite the year old circuit decision in Neville. He acknowledged that Griffith itself strongly suggested that interest analysis would be utilized in contract cases by its approval of a leading New York contract choice case, Auten v. Auten, 308 N.Y. 155, 124 N.E.2d 99 (1954). But he also noted the reassertion of traditional rules in Crawford. Finally, he argued that Hunter's oblique suggestion that "interest analysis" applied to contracts could be explained away as "only tangentially present[ing] a conflicts of laws problem . . .." 437 F.2d at 529, n. 6 (and see note 45 supra). Judge Aldisert concluded that:

Pennsylvania's highest court has not so clearly embraced the grouping-of-contacts theory as to set aside a choice of law provision explicitly set forth in the written agreement which is the subject matter of the dispute.

437 F.2d at 530. Judge Aldisert, however, did not have to determine exactly what Pennsylvania's methodology was nor apply that methodology since in Boase the parties had explicitly chosen a law to govern the transaction; his inquiry was limited to the question whether Griffith and its progeny would require not giving effect to the parties private choice. He concluded Griffith would permit giving it effect. Boase was followed shortly by Pittsburgh Bridge, where the court opined tangentially that Pennsylvania retained the traditional reference to the law of the state where an insurance contract is made ("issued and delivered"), citing only Varas and Eastcoast Equipment.

The Pennsylvania Supreme Court decision in Danz, applying lex loci contractus to a contracts case, followed. As we mentioned, Danz cited neither Griffith nor its progeny. The questions naturally arose whether Pennsylvania was reviving the "vested rights" approach generally, whether it was doing so but only in the area of contracts, whether Pennsylvania intended to apply different choice of law methods for inter-state and international cases, or whether Danz was merely an aberration. Given the potential significance of Danz to the evolution of Pennsylvania choice of law, it is surprising that it has received so little notice in the Circuit.

Travelers, which followed Danz by several years, suggested that Pennsylvania would use interest analysis in contracts choice of law. It did not cite Danz, but rather relied solely on Griffith in holding that Pennsylvania would apply interest analysis to a case concerning arbitrability of the meaning of an insurance contract; Massachusetts was found to be "interested" because "the contract was made there," but Pennsylvania's interest was greater because "decedent and his executors are citizens of that state." 490 F.2d at 543. In Siata International, an action on a bond covering a contract, the court relied on Neville in finding that the district judge had employed the correct test of "most significant relationship," but remanded for a determination, inter alia, of whether an express choice by the parties should be considered as an additional factor in the calculus, noting that in the process "additional state interests and contacts may become manifest." 498 F.2d at 820.

Kademenos, a case of tortious interference with contractual relations, hence straddling between tort and contract, was the next Circuit case in the wake of Danz. It cited Griffith and Neville and spoke of Pennsylvania's "greater interest," but in analyzing that interest utilized Restatement II territorial contacts exactly as Neville had, focusing on the place of the policy's delivery, the place of the policy's cancellation, etc. 513 F.2d at 1073. Daburlos, however, which also followed shortly thereafter returned to Restatement I language in saying that:

[i]nitially, there can be no question that under the doctrine of conflict of laws, the Pennsylvania substantive law applies for the insurance policies in this case were issued and delivered in Pennsylvania.

521 F.2d at 20. The Court's footnote somewhat cryptically cited Griffith, Pittsburgh Bridge and Neville.

We now turn to the Circuit cases which followed the most recent development in the Pennsylvania state cases, Gillan, in which the Superior Court for the first time read Griffith expansively (while ignoring Danz), and in which that court seemed to prescribe Restatement II analysis for contract cases. Gillan, of course, left a question whether it was to be read as limited to contracts arising out of marital dissolutions. Three Third Circuit cases have followed Gillan. The first two cases, Craftmark Homes and William B. Tanner, came down only two months after Gillan, doubtless before it was generally reported, and did not discuss it. Craftmark cited only Danz's "place where a contract is made and where it is to be performed" language as dispositive of the facts in the case. 526 F.2d at 792 note 2. William B. Tanner, an action for anticipatory breach of contract, found no threshold conflict between Pennsylvania and Tennessee law, hence no need to apply choice of law principles, but found that

[f]urthermore Pennsylvania is the place of performance under the contracts. Although the point is not free from doubt, we interpret Pennsylvania law as providing that ``matters connected with the performance of a contract are governed by the law prevailing at the place of performance.' Musser v. Stauffer, 192 Pa. 398, 43 A. 1018 (1899).

528 F.2d at 266. This passage is, of course, consistent with Restatement I. The final Circuit case, Jamison, is similarly unhelpful in trying to reconcile or distinguish the state cases, for the Jamison panel cited only Crawford in holding that the interpretation of an insurance contract is governed by the law of the place where it was made. 536 F.2d at 562 n.1.^[50]

A district judge must be guided in the first instance by circuit law. As the foregoing discussion demonstrates, circuit law does not guide us here, although the responsibility for the "drift" in this area must, as we have already noted, be laid at the doorstep of the Pennsylvania Supreme Court not the Third Circuit. We must therefore advance our discussion on the basis of our independent analysis of the Pennsylvania cases.

3. A Disquieting Verdict

In Griffith, its first modern choice of law opinion, the Pennsylvania Supreme Court avowed its intent to "re-examine its position" in the field of choice of law. Since the purpose of the proposed re-examination was to bring the state's "rules" more into "harmony with the realities of this age," 416 Pa. at 11, 203 A.2d at 801, the Court did not limit the scope of its vow to any particular aspect of choice of law — e. g., torts as opposed to contract. Despite the difficulties attendant upon such a broad revision of an area of law, the Pennsylvania Supreme Court insisted — doubtless too optimistically — that its new approach would

not be chaotic and anti-rational. ``The alternative to a hard and fast system of doctrinal formulae is not anarchy. The difference is not between a system and no system, but between two systems . . ..'

416 Pa. at 22, 203 A.2d at 806, citing Harper, Policy Bases of the Conflict of Laws: Reflections on Rereading Professor Lorenzen's Essays, 56 Yale L.J. 1155, 1157-1158 (1947).

As we have seen, that view proved less than prescient. For: (1) some recent state cases have applied the old rules to contract cases with barely a nod to Griffith (Eastcoast Equipment, Crawford, and Danz); (2) other state cases have extended Griffith beyond the tort context of its birth while neither over-ruling nor explaining the cases which continued to use the old rules (Hunter, Gillan); (3) the post-Griffith tort cases have moved from pure interest analysis into a hybrid interest analysis/Restatement II approach with strong territorial overtones, a system which turns to territoriality to resolve the choice when (and if) interest analysis is applied and fails; and (4) the Third Circuit has developed a line of traditional contract choice of law cases and one of modernist contract choice of law cases without acknowledging their inconsistency. Thus it seems fair to conclude that Pennsylvania has not "a" system of choice of law but two if not three competing systems, each claiming the same legal territory and leading to enormous confusion.

Put differently, our reading of the Pennsylvania case law and of the Third Circuit decisions have led to incompatible inferences: that lex loci contractus is still viable in Pennsylvania, and that it has been abandoned in favor of a Griffith-type inquiry.

Under these circumstances we deem it appropriate to apply Restatement I and the interest analysis/Restatement II hybrid as alternative methodologies in solving our choice of law problem. It is to an application of those methodologies that we now turn after a brief but necessary discussion of the parties' contentions and of the reasons for our original bench opinion.

F. The Parties' Contentions; The Impact of Van Dusen v. Barrack

We have not yet formally addressed the parties' choice of law contentions. In view of the length of this opinion we will recite those contentions and our reasons for rejecting them briefly, dwelling extensively only upon one contention of the plaintiff. Plaintiff suggested three paths to reach her desired result of the choice of New York law on the presumption against suicide. First, she contended that Griffith's interest analysis revolution applied only to torts and not to contracts, that the traditional lex loci contractus rule remained the rule under Crawford, and that under Crawford the state law controlling construction of the contract is that of New York, because that is where the last act necessary to the making of the contract occurred. Plaintiff, of course, recognized that the issues of presumption and burden of proof often have been treated as matters of procedure rather than of substance; nonetheless, she also contended that Pennsylvania would, on those issues, follow Restatement II, §§ 133, 134 in referring to the law of "the state of the otherwise applicable law [if the purpose of that law] is to affect the decision of the issue. . . ." We were unwilling to accept this argument when it was offered because plaintiff offered us no reason to believe that Pennsylvania would choose that particular mixture of theories; i. e., Restatement I as to contract construction and Restatement II as to presumption and burden of proof in the manner plaintiff then had posed them. In fact, plaintiff did not explain, but simply took advantage of the confusion among the Pennsylvania cases, e. g. she did not explain or consider the impact of Gillan v. Danz on the contract issue. Nor did she justify rejecting, modifying, or limiting Pennsylvania's lex fori rule for burden/presumption questions. In sum, plaintiff's convenient selections among the authorities seemed to us arbitrary and unconvincing. And our own analysis had not then advanced to the point it now has.

Next, plaintiff insisted that if we were to hold that Griffith had changed Pennsylvania choice of law in (insurance) contract cases, we should still apply New York law because New York is the state with the dominating interest. She supported that contention mainly on the grounds that the insurance policy conforms to — indeed uses the very language of — the New York Insurance Code, and that defendant is a New York corporation with its principal place of business in New York. In fact, however, the coincidence in language between the policy and the New York Insurance Code is matched by their mutual coincidence with the Pennsylvania and Delaware statutes. Compare New York Insurance Law § 164 (McKinney 1966) with Pa.Stat.Ann. tit. 40, § 753 (1971) and Del.Code Ann., tit. 18 §§ 3303 et seq. (1975). Also, the "interests" which Griffith and its progeny found determinative were the interests which states have in protecting their domiciliaries. Plaintiff offered no basis in Griffith for finding that a state (i. e. New York where the defendant is incorporated) is interested in imposing duties or burdens on its citizens in their multi-state contacts which are imposed neither by the locus of the activity nor by the state of domicile of the other party.

Finally, plaintiff offered us an intriguing and appealing way around the morass of Pennsylvania choice of law; she suggested that we treat the New York State Court's forum non conveniens dismissal as equivalent to a federal interdistrict transfer under 28 U.S.C. § 1404(a), which permits transfer of cases between federal judicial districts "[f]or the convenience of parties and witnesses, in the interest of justice." In Van Dusen v. Barrack, 376 U.S. 612, 84 S.Ct. 805, 11 L.Ed.2d 945 (1964), the Supreme Court held that § 1404(a) does not affect the applicable law and that, as an implication of Erie, the transferee court must apply the transferor court's choice of law rules or methods. If plaintiff were to prevail on this point we would be obliged to apply New York instead of Pennsylvania choice of law, which plaintiff believed would be more likely to result in application of the New York presumption against suicide.

In contending for the application of the Van Dusen rule, plaintiff argued that since § 1404(a) was a codification of state forum non conveniens rules, the incidents of the offspring should be projected back onto the parent. She did not articulate whether she believed we could so project as a matter of Pennsylvania law adopted under Erie or as a matter of federal policy. However, neither state nor federal law compels such a decision.^[51] No state has adopted a rule that when a foreign court has dismissed a case on forum non conveniens grounds, it will adopt the choice of law rules of the foreign court for use in that case. Nor has plaintiff proffered any reason to convince us that Pennsylvania or any other state would adopt such a rule.^[52]

But if the states have not adopted and will not adopt that rule, then the federal Erie-Van Dusen policy of

ensur[ing] that the ``accident' of federal diversity jurisdiction does not enable a party to utilize a transfer to achieve a result in federal court which could not have been achieved in the courts of the state where the action was filed,

Van Dusen, 376 U.S. at 638, 84 S.Ct. at 820, could not be served even by extending the Van Dusen "rule" to the present context. Indeed, such an extension would "ensure" a different decision as to whose choice of law rules applied as between the state and federal courts of the convenient forum; i. e., if after a forum non conveniens dismissal in one state court, plaintiff sues in a "convenient" state court, the latter state will apply its own choice rules; but if plaintiff brought the second suit instead in the federal court of the same "convenient" forum and if Van Dusen were to be extended to that context, then the federal court would apply the original forum's choice of law. This conflicts with the Erie-Van Dusen policy of avoiding different results in the state and federal courts of the same state.^[53]

Additionally, we note that the Van Dusen court stressed that it was engaged in statutory construction of § 1404(a). And finally, the criteria for § 1404(a) transfer are different from those for state forum non conveniens. See Parsons v. Chesapeake & Ohio R. Co., 375 U.S. 71, 84 S.Ct. 185, 11 L.Ed.2d 137 (1963). For all the foregoing reasons, we do not believe that Van Dusen is applicable as a matter either of state or of federal law.

We turn now to defendant's contentions, which we found equally unconvincing. First, defendant argued that Griffith established the interest or "significant contacts"^[54] approach as the sole guideline to Pennsylvania choice of law and that Delaware was the proper selection under that (those) approach(es). Defendant's total reliance on Griffith was hardly helpful in light of its failure to explain the post-Griffith non-interest cases and Danz. Additionally, we note that defendant did not point to any interest of Pennsylvania or of Delaware in having a heavier burden of proof placed on the beneficiary of the accident policy than the burden placed thereon by New York.^[55]

Second, defendant proposed that it made no difference which methodology Pennsylvania applied because the traditional approach led to application of Pennsylvania law on presumption/burden as lex fori. We have dealt with (and rejected) this contention in part IV.(B), supra.

G. Our Pretrial Bench Opinion

We initially addressed the contentions of the parties in a pretrial bench opinion. It is therefore important that we briefly describe that opinion and our reasons for not following its rationale now (though we still subscribe to its result).

Before trial we concluded that Crawford was good law on the narrow issue of choice of law for insurance contract construction. Our further research has led us to qualify our reasoning. Then, and now, as will be seen, we believe that New York law controls contract construction. But our bench opinion was mainly devoted to the question whether or not New York law also controlled on the presumption issue. We suggested, then, that that issue did not require us to predict Pennsylvania choice of law, but rather only to apply Fed.R.Evid. 302 which says

the effect of a presumption respecting a fact which is an element of a claim or defense as to which state law supplies the rule of decision is determined in accordance with state law.

We had no doubt that the "presumption" against suicide fell under that rule as one affecting an element of a state law claim or defense; we questioned only which state's law was to be applied in our conflicts situation. In resolving that issue we relied in large measure on Pilot Life Insurance Co. v. Boone, see discussion p. 1078 supra, in which Judge Rives reversed a district court decision which had applied forum law qua forum law to the presumption against suicide and in which he had classified the presumption as a matter of substantive contract law because it is inseparably connected with the right to defend. Based on the Pilot Life analogy, we concluded that Rule 302 referred us to the law of that state whose law supplied the rule of decision, i. e., New York law under Crawford.

Upon further reflection, we realize that such a reading of Rule 302 might be taken as having put the rabbit in the hat as it were by short circuiting or circumventing the Klaxon -mandated Pennsylvania choice of law analysis, and therefore as conflicting with Erie/Klaxon, which the rule cannot be deemed to override. And while the Advisory Committee note to Rule 302 cited, inter alia, Palmer v. Hoffman, 318 U.S. 109, 63 S.Ct. 477, 87 L.Ed. 645 (1943), a diversity choice of law case in which a federal district court had applied a foreign state's burden of proof rule, the Supreme Court upheld that decision not on the ground that there is a special federal rule on choice of burden of proof, but on the ground that the forum state would so refer to the foreign rules. See also Weber v. Continental Casualty Co., 379 F.2d 729 (10th Cir. 1967). We now conclude that Rule 302 cannot be read as solving our problem by creating a federal choice of law rule at odds with the state choice of law rules referred to under Erie-Klaxon.

Having explained why we cannot follow the reasoning of our bench opinion now, we turn to an application of our conclusions about the present state of Pennsylvania choice of law to this case.

H. Application of Pennsylvania Choice of Law Principles to this Case

1. Analysis Under Restatement I and Lex Loci (Contractus) Rules

If we infer from Danz, Eastcoast Equipment, and Crawford that Varas v. Crown Life Insurance Company is still good law, then we must apply the traditional territorial rules of Restatement I. (See p. 1094 supra). The underpinnings of Restatement I are its purported certainty, predictability, and ease of application. Whatever may be said about certainty and predictability, ease of application does not shine through here. At the threshold, Restatement I was not designed to solve our problem. Inasmuch as all presumptions are considered "procedural" by virtue of § 595(2), and the forum automatically applies its law to all procedural matters, no conflict of law problem would arise for a court rigorously adhering to Restatement I. However, we have decided the Pennsylvania courts would not follow § 595(2), and would treat the presumption of non-suicide as "substantive," necessitating application of choice of law principles. Thus we must attempt to use the substantive body of the Restatement, discerning the closest analogues to the presumption against suicide in a life insurance contract.

Our first decision is whether the presumption is a matter of "contracting" or a matter of "performance." If the former, questions will be governed by § 311 on the "place of contracting;" if the latter, § 355 on the "place of performance" will govern.^[56] Our presumption seems to be both and neither. On the one hand, the presumption relates to the "obligations" of a contract, § 346, which is determined by the law of the place of contracting. The comment to § 346 provides:

whether a duty of performance is to be conditional, and when, where and by whom the performance is to be made and exactly what act of performance a party is obliged by the contract to do are determined by the law of the place of contracting.

Similarly, § 347 on defenses seems applicable since it points to the place of contracting for the issue of "any legal or equitable defense."

On the other hand, § 359 tells us that the "place of performance" governs the "performance or happening of a condition." Since "accidental" death is a condition precedent to recovery under the policy, presumptions relating thereto would arguably be a function of the place of performance. Additionally, § 370 tells us "the law of the place of performance determines whether a breach has occurred." Suicide could be construed as breach, giving rise to non-liability by the insurer. Therefore two sections point to the law of the place of contracting and two to the place of performance. All four provisions are arguably applicable, but none fits precisely.

If one and the same state is both the place of contracting and the place of performance, then we need not make the above choice. Therefore, we turn to the appropriate Restatement I sections to determine exactly where the "place of contracting" and the "place of performance" are located. Three sections, 317, 319 and 326 are potentially applicable for place of contracting, and one section, 355, is applicable to place of performance.

In the case at bar, the sequence of events leading to the purchase of the insurance policy was as follows: Josiah Marvel Scott contacted the Delaware branch office of an independent insurance broker with principal offices in Philadelphia. The Philadelphia office called the insurance company in New York, and effected an oral binder. Subsequently, a written application was transmitted from Philadelphia to New York. The written policy was issued from New York by mail to the broker in Philadelphia, who transmitted it to the Delaware office, who in turn delivered it to the insured. See Pretrial Order, paragraphs 10-14.^[57]Restatement I § 317 provides that when, as here, the policy is effective upon posting by the insurance company — i. e. no further acts are required by either the broker or the insured to effect coverage — the place of posting, here New York, is the place of contracting. Similarly, § 319 provides:

When an offer for an insurance company contract is received by the company through a broker who acts for a client, and the policy is effective on delivery, the place of contracting is where the policy is posted or otherwise delivered to the broker.

This section is also applicable to the case at bar, since the broker was clearly acting for the client (in distinction to § 318 where the broker is in reality merely an agent for the insurance company). Like § 317, § 319 points to the law of the place of posting the final written policy, which was New York. Finally, § 326(b) provides:

When an offer for a bilateral contract is made in one state and an acceptance is sent from another state to the first state in an authorized manner the place of contracting is as follows:

. . . . .

(b) if the acceptance is sent by any other means, the place of contracting is the state from which the acceptance is sent.

Section 326(b) is applicable to the creation of the oral binder. The "offer" was made in Philadelphia, and the "acceptance" occurred in New York. Pursuant to § 326(b) the "any other means," which includes the telephone call, points to New York as the place of acceptance, and hence the place of contracting. Indeed, the comment to this section provides that "when an acceptance is to be given by telephone, the place of contracting is where the acceptor speaks his acceptance." In sum, all three of the possibly applicable sections point to New York as the place of contracting.

Turning to the place of performance, § 355 is of little help, providing with circularity that "the place of performance is the state where . . . the promise is to be performed." Since, in every imaginable sense, the insurance company would have "performed" its obligation fully under the contract by writing and posting a check to beneficiary, which action would have occurred in New York, we find that New York is also the place of performance. Since New York satisfies both places, of contracting and of performance, we need not decide whether the presumption is a matter of contracting or performance. We therefore conclude that if lex loci contractus is still viable in Pennsylvania, a Pennsylvania court would determine New York's law including its presumption against suicide to be applicable.

2. Post-Griffith Interest/Restatement II Hybrid Analysis

In the alternative, we may infer from Justice Roberts opinions in Griffith, McSwain, Elston, Hunter, and Kuchinic, and from Cipolla and Gillan, that lex loci contractus has been eroded, if not abandoned, in Pennsylvania. (See p. 1094 supra). In that event, we must apply the "interest analysis/Restatement II hybrid" method we discerned in these cases. In each case, the starting point has been interest analysis.

(a) Interest Analysis; Lex Loci v. Lex Fori in the "Unprovided for Case."

We previously described New York's law in terms of the "fair question" doctrine, and Pennsylvania's (and probably Delaware's) law as the "Watkins" doctrine. See pp. 1081-1082 supra. We begin by considering what interests either of these latter two states might have in applying Watkins in a multistate context. We note, first, that Watkins could serve the interest of a state in protecting domestic insurance companies from "fraudulent" claims. See Babcock v. Jackson, 12 N.Y.2d 473, 240 N.Y.S.2d 743, 191 N.E.2d 279 (1963). But the defendant here is a New York corporation, and there is no basis in this record for us to infer that either Pennsylvania or Delaware would have any interest in protecting the defendant from obligations imposed by the rules of its home state.^[58]

Second, either state might prefer to have its domestic law applied to a multi-state transaction in order to keep down insurance rates for its citizens. See Cipolla v. Shaposka, supra. But see Morris, Enterprise Liability and the Actuarial Process — The Insignificance of Foresight, 70 Yale L.J. 554 (1961) (asserting that local liability law variation has negligible effect upon automobile insurance rates). Yet it is clear that Pennsylvania's interest in lowering insurance rates for its citizens could not be promoted by application of its law to an insurance policy covering the risk of accidental death of a non-citizen. The policy did, of course, insure a citizen of Delaware; but we note parenthetically that there is nothing in the record which demonstrated whether defendant's premium rate varied with residence,^[59] We note also that Delaware asserts its legislative control only over those who "transact" insurance in that state. Del. Code tit. 18, § 505 (1975).^[60] But, the evidence before us indicates no more contact between the defendant and Delaware than the mere fact that the defendant accepted plaintiff's offer to purchase a single insurance policy. Cf. Aldens, Inc. v. Packel, 524 F.2d 38, 43 (3d Cir. 1975) (leaving open the question of whether attempt to "exercise . . . the state's sovereign decisional authority" on account of such meager contact violates due process.). Finding no other cognizable interests, we conclude that defendant has not demonstrated that the application of New York's presumption against suicide needlessly frustrates any interest of Pennsylvania or Delaware in applying its own law to this case.

Turning the coin, we note that no New York interest would have been frustrated had we applied the contrary rules of Pennsylvania or Delaware. In interest analysis terms, the New York rule appears to be intended to protect the interest of a New York insured in providing for his survivors, and the interest of a New York beneficiary in being able to collect. Here neither the insured nor the beneficiary was a citizen of New York. And New York has no interest in imposing more burdensome obligations on its own citizens (including the defendant) in their dealings with foreign citizens than those which are imposed by foreign states on their own domestic transactions.^[61]

We have before us, then, a specimen of that rare species which Professor Currie has dubbed the "unprovided [for] case"; i. e., a case in which "[n]either state cares what happens." Currie, Selected Essays on the Conflict of Law 152 (1963). See p. 1078 supra. So rarely have such cases been recognized that Currie had to turn to pre-interest cases to exemplify the fact pattern he meant.^[62] In recent years a few such cases have been decided; they will be discussed below. Professor Currie himself expected "[t]raditionalists [to] stand aghast at this anomaly, and [to] take it as proof of the unsoundness of the analysis." He insisted that it was not so much a defect as a challenge to consider on what terms a judge should decide the "no interest" cases.

Judging that we should "concern . . ourselves about the results," Currie rejected out of hand any notion that in the absence of an interest a court should apply the traditional mechanistic rules. Currie, supra, at 153. Indeed it was at this point in his argument that Currie suggested that lex fori rather than lex loci should be applied to any issue in the absence of a sufficiently compelling foreign sovereign interest. His sole argument was that of "convenience" or efficiency: this is the rational and convenient way to try a lawsuit when no good purpose is to be served by putting the parties to the expense and the court to the trouble of ascertaining the foreign law.

Id. at 156.

Currie's solution has received mixed reviews in those few judicial opinions dealing with the problem of choice of law in "unprovided-for cases." Thus, in Erwin v. Thomas, 264 Or. 454, 506 P.2d 494 (1973), the wife of a Washington domiciliary who had been injured in Washington by a resident of Oregon brought suit in Oregon for loss of consortium. In domestic Oregon cases such loss was actionable; in domestic Washington cases it was not. In the multistate context neither state's interest in protecting its domiciliary would have been served by applying its law. In that situation, the Oregon Supreme Court adopted Currie's residual lex fori rule on the sole ground of case or convenience:

neither state has a vital interest in the outcome of this litigation and there can be no conceivable material conflict of policies or interests if an Oregon court does what comes naturally and applies Oregon law.

Id. at 459, 506 P.2d at 497.^[63]

New York took the opposite tack in Neumeier v. Kuehner, 31 N.Y.2d 121, 335 N.Y. S.2d 64, 286 N.E.2d 454 (1972), an unprovided for case in the mold of Erwin and Labree (see note 63 supra). Instead of applying lex fori in the absence of an "interested" state, the New York Court of Appeals applied lex loci (delictus). It hesitantly proposed the policy/interest distinction (whose significance we just noted, see note 60 supra) in the following terms:

although New York has a deep interest in protecting its own residents, injured in a foreign state, against unfair or anachronistic statutes of that state, it has no legitimate interest in ignoring the public policy of a foreign jurisdiction . . . and in protecting the plaintiff guest domiciled and injured there from legislation obviously addressed, at the very least, to a resident riding in a vehicle traveling within its borders.

31 N.Y.2d at 125, 335 N.Y.S.2d at 68, 286 N.E.2d at 456 (emphasis supplied). This language suggests to us that Chief Judge Fuld saw the territorial sovereign as having policies that are always appropriately "addressed" to a state's residents while they are within its territory. It is only when another state's "interest" in protecting its domiciliary clashes with such a general "policy" that the exclusivity of territoriality is to be questioned and the conflicts (or false conflicts) of interest are to be examined.

Although Judge Fuld did not develop further the theoretical significance of the policy/interest notions, he did suggest a compelling practical reason to prefer lex loci in the absence of a demonstration that such would needlessly frustrate another sovereign's interest; namely, to prevent forum shopping. We agree that as a general rule, the adoption of a rule of lex fori rather than one of lex loci would promote forum shopping. For example, in Erwin, plaintiff won only because she sued in defendant's home state. Had she sued in her own home state — the state in which the accident occurred — and had Washington also adopted the lex fori rule for residual cases, she would have lost. Such forum shopping could also lead plaintiff to prefer suing in a federal court, because if the forum were found to be inconvenient, plaintiff could still get the benefit of the law of the selected forum. See Van Dusen v. Barrack, supra. Such possibilities do not commend lex fori.

In sum, the unprovided-for case challenges us to consider both the structure of choice of law and the meaning of interest analysis. The contending proposals have not fully plumbed that problem in terms of this case. Currie justified his solution, applying lex fori, whenever no other sovereign has been shown to have a compelling interest, only on grounds of efficiency or convenience. Doubtless the inconvenience of applying foreign law can be great; indeed, in some cases the difficulty of proving foreign law may be insuperable or unreasonably expensive. See Walton v. Arabian American Oil Co., 233 F.2d 541 (2d Cir.) cert. denied, 352 U.S. 872, 77 S.Ct. 97, 1 L.Ed.2d 77 (1956). Yet the strength of Currie's position is sapped when it is tested against Neumeier's alternative proposal that, in the absence of proof that another sovereign has a compelling interest in having its law applied to a particular issue, lex loci should be applied. Neumeier's argument that such a rule would prevent or reduce forum shopping seems at least as substantial as the "convenience" argument.

Moreover, had Neumeier more amply developed its forum shopping contention, it might have pointed out that the very judicial economy promised by the advocates of lex fori (because the lawyers and the judges of the forum can more easily ascertain their local law) could be undercut whenever the parties who stood to benefit from the differing laws of the several states could countersue in the states of their choice. The ensuing race to judgment would be neither convenient, nor efficient, nor seemly. Cf. Hanson v. Denckla, 357 U.S. 235, 78 S.Ct. 1228, 2 L.Ed.2d 1283 (1958).

Just as Neumeier rejected application of lex fori in the unprovided-for situation, we believe that Pennsylvania would also reject its automatic application. In the first place, the "unprovided-for" case is the mirror-image of the true conflict situation: in the latter both jurisdictions are interested, in the former no jurisdiction is. Professor Currie treated them as mirror-images, advocating the application of lex fori in both instances. The Pennsylvania courts have never been presented with an unprovided-for case. But on the two occasions where they have discerned true conflicts — Elston and Cipolla — they have eschewed lex fori and deferred to the other state's law because of territorial contacts.^[64] We conclude that were the Pennsylvania courts faced with a case where an initial application of interest analysis demonstrated no interested jurisdictions—in other words, the "unprovided-for" case presently before us — they too would treat it as the mirror-image of Elston and Cipolla. And just as they did in those two cases, we believe they would reject lex fori and seek a solution through territorial contacts, for that has been the consistent theme of the Pennsylvania "hybrid" system.

Furthermore, we conclude there would be substantial justification for this choice in Pennsylvania policy. For the values served by lex fori in the true conflict or unprovided-for case are those of convenience because the lawyers and the judges of the forum can more easily ascertain their local law. In contrast, reliance on territorial contacts promotes an even greater policy, the prevention of forum shopping. We note Pennsylvania's long tradition of discouraging forum shopping,^[65] and conclude that that important value would be served by rejecting lex fori.

(b) Territorial Contacts Analysis

We have concluded Pennsylvania courts in the unprovided-for case would turn to the other part of its "hybrid" system, territorial contacts.^[66] This is the method of Elston and Cipolla in true conflicts, and Griffith, McSwain, Hunter and Kuchinic in false conflicts. Therefore, it is to such territorial contacts that we turn (alternatively) in search of a solution to our problem.

The threshold question, crucial to our analysis, is which territorial contacts would be deemed important. Since the hybrid system has not as yet been employed by the Pennsylvania courts in a contracts case, we must reason by analogy from the tort context.

One possibility would be to adopt the specific jurisdiction-selecting rule of § 192 of Restatement II. That provides:

The validity of a life insurance contract issued to the insured upon his application and the rights created thereby are determined, in the absence of an effective choice of law by the insured in his application, by the local law of the state where the insured was domiciled at the time the policy was applied for, unless, with respect to the particular issue, some other state has a more significant relationship under the principles stated in § 6 to the transaction and the parties, in which event the local law of the other state will be applied.

Were this provision to be adopted, it would point to the law of Delaware, where Jay Scott was domiciled at the time of the policy's issuance.

It seems to us highly unlikely a Pennsylvania court would adopt or rely on § 192. In the first place, the post-Griffith "hybrid" case law demonstrates the method of examining all the territorial contacts, not of just isolating one contact and making it dispositive. For example, Griffith examined where the relationship had begun, Kuchinic examined the parties' domiciles as well as where the relationship began and was to end, and Cipolla looked at the place where the relationship began, the place where the accident occurred, and the place where the car was registered. If we may extrapolate from this, it seems apparent that in a tort context, the hybrid system looks to all the territorial contacts made relevant by § 145, the general tort principle. These include place where the injury occurred, place where the conduct occurred, domicile, residence and nationality of the parties, and the place where the relationship between the parties was centered. Concomitantly, were the Pennsylvania courts to apply the hybrid system to a contracts case, they would almost surely apply § 188, the general principle for contracts, and the counterpart of § 145. A modicum of support for our conclusion stems from Gillan, which considered a separation agreement, the closest the Pennsylvania courts have come to a contract action. Gillan set out the entire § 188 as framework for its discussion, 236 Pa.Super. 147, 151, 345 A.2d 742, 744 n.1.

In the second place, choosing a specific jurisdiction-selecting rule like § 192 would run counter to the entire Griffith hybrid system. After all, the point of Griffith was to overthrow the ancient system which gave dispositive weight to one isolated factor ("the wooden application of a few overly simple rules, based on the outmoded vested rights theory"). It would be wholly illogical for a court attempting to apply the Griffith methodology to a contract action to embrace the very type of rule Griffith was committed to abandoning.^[67] We therefore conclude all relevant territorial contacts under § 188 would be examined.

Those factors are: (a) the place of contracting; (b) the place of negotiation of the contract; (c) the place of performance; (d) the location of the subject matter of the contract; and (e) the domicile, residence, nationality, place of incorporation and place of business of the parties. We will consider these factors seriatim. "Place of contracting" is meant, by comment on subsection (2) to describe the place "where occurred the last act necessary . . . to give the contract binding effect." As stated earlier, since an oral binder was established by telephone and the written policy was subsequently posted from New York, New York is unequivocably the place of contracting. The "place of negotiation" is ambiguous. It occurred in Delaware, where the broker took Scott's application for the contract, in Pennsylvania, where the broker's main office contacted the insurance company by placing the call and then posting the written application, and in New York, where the insurance company agreed to the terms, both orally and subsequently in writing. In any event, the applicable comment makes it clear that this factor diminishes in importance "when the parties do not meet but rather conduct their negotiations from separate states by mail or telephone."

The third factor is "place of performance." The comment does not elucidate this contact, and we conclude, as with the identical language of Restatement I, that this most fairly refers us to New York, where the insurance company would have fully performed its obligation by posting a check to the beneficiary. The fourth factor is "situs of the subject matter of the contract." The comment states that "when the contract deals with a specific physical thing, such as land or a chattel, or affords protection against a localized risk, such as dishonesty of an employee in a fixed place of employment, the location of the thing or of the risk is significant." (Emphasis added.) The risk insured against, of course, was Scott's accidental death. But this was not a localized risk, since the policy by its terms covered Scott virtually everywhere. By negative inference from the above comment, the situs is not significant when the risk is not localized. The fifth factor includes the domicile, residence, place of incorporation and place of business of the parties. Plaintiff was a Pennsylvania resident at time of suit, and defendant was both incorporated and had its principal place of business in New York.

Considering these factors, New York's contacts, while not overwhelming, appear definitely more substantial than Delaware's or Pennsylvania's. New York has the exclusive contact for two of the five factors, shares another with Pennsylvania and another with both Pennsylvania and Delaware. It's only "non-contact" is with the situs, but the situs seems to have small or non-existent significance in our case of a non-localized risk. In contrast, Pennsylvania has only the two shared contacts of negotiation and party residence, and Delaware has a weak, if not non-existent contact in the "situs" and a shared contact in the negotiation. In sum, we conclude that the balance of territorial factors point to New York's as the applicable law under the post-Griffith hybrid methodology.

It remains to be considered whether there are other relevant territorial contacts besides those factors delimited in § 188. Specifically, Delaware at an intuitive level seems to have numerous territorial connections: the plane crashed there, a good deal of the family fortune which Jay Scott squandered (an alleged motive for suicide) was located there, and so on. On analysis, we conclude these factors have little relevance. They would be relevant if we had before us a motion to dismiss on forum non conveniens grounds — i. e. because many of the witnesses were in Delaware. But choice of law and forum non conveniens considerations are very distinct. What is before us is a suit by a Pennsylvania beneficiary against a New York insurance company, and the Delaware "contacts" on the crash and motive for suicide are not relevant to the choice of law question. Even if they were we cannot conclude they would outweigh the New York contacts delineated above in considering § 188.

3. Conclusion

Both the traditional rules (Restatement I) and the hybrid interest analysis/Restatement II approaches used by the Pennsylvania courts have led us to the conclusion that New York's "fair question" doctrine on the presumption of accident is applicable. This is the conclusion we originally reached in our pretrial bench opinion, albeit on what we now consider untenable grounds (See IV(G) supra). Therefore, we reaffirm our decision that New York's law on the presumption was the proper one to have chosen to govern this trial.

Our conclusion that both Restatement I and the post-Griffith hybrid approach point to the same state's law make it unnecessary to decide whether lex loci contractus as a choice of law principle has been abandoned by the Pennsylvania courts.^[68] We whole-heartedly encourage them to make this decision. It is difficult enough to apply a state's choice of law methodology rationally and faithfully; it is confusion compounded when it is not initially clear which methodology is appropriate. If this lengthy opinion has served no other purpose, perhaps it will impress upon the Pennsylvania Supreme Court the problems which its failure to clarify its choice of law contract rules have created for nisi prius judges, state and federal, and for the United States (Third Circuit) Court of Appeals. Needless to say, we earnestly hope that the Pennsylvania Supreme Court, at its first opportunity, will resolve the confusion in its choice of law jurisprudence which we have labored so long to relate and for purposes of this case, resolve.

V. The Charge to the Jury

Earlier (p. 1082 supra), we briefly described New York's presumption against suicide as it is set forth in the Wellisch and Begley decisions. Having determined that this was the proper presumption to have chosen under Pennsylvania choice of law rules, we must now consider whether the charge given the jury accurately reflects New York law. The relevant portion of our charge was:

If after considering all the evidence, there is a fair question in your mind as to whether the crash was due to an accident or to suicide then because of the presumption against suicide your verdict should be for the plaintiff. If, on the other hand, after considering all the evidence there is no fair question in your mind but that Mr. Scott committed suicide, then your verdict should be for the defendant.

Earlier, we instructed,

[if] after considering all the evidence the jury finds that a fair question of fact is presented as to whether death was due to suicide or accident then the jury should answer accident.

Charge of the court at 20, 21. Defendant insists that we mistook New York law by treating the existence of a "fair question" as an issue of fact to be determined by the jury, because, defendant says, the New York Court of Appeals has used the "fair question" language to define a question of law for the court. Correspondingly, defendant argues that we erred in construing the New York cases as holding that if the jury does find there to be a fair question, then they must give a verdict for the plaintiff, when, in fact, those cases held only that where the court finds a fair question to exist, then the court must permit the case to go to the jury.

We expressed earlier our view that Wellisch and Begley did not precisely clarify the effect of the presumption against suicide. It is necessary to give meaning as best we can to their words.

Wellisch, written in 1944, was a suit on a life insurance contract excepting suicide. Accordingly, defendant insurance company had the burden of proof, and the appeal concerned defendant's motion that it should have been granted a directed verdict. The court examined the evidence to determine its sufficiency to support the jury's finding, and having determined a jury question to be clearly present, went on to add:

That presumption is not one of those that takes the place of evidence so as to create a question of fact even when all the real proof is the other way . . . Nor is it the sort of "presumption" that serves only to shift the burden of proof and disappears from the case as soon as evidence to the contrary is offered. It is really a rule or guide for the jury in coming to a conclusion on the evidence. The "presumption against suicide" means that when death by violence is shown and an inference must be drawn by the jury as to suicide or not, then the jury should in justice and good conscience draw the inference of accident, not suicide. This presumption is one aspect of the broader rule that where evidence is susceptible of two constructions, the construction which does not imply criminality or moral turpitude is to be favored. . . . It means . . . that a fair question of fact as to accident or suicide should be answered: "accident." (Emphasis added.)

293 N.Y. 178, 184, 56 N.E.2d 540, 543.

Twelve years later Begley applied this language to a double indemnity provision which, like the accident policy in our case, initially put the burden on plaintiff beneficiary to plead and prove accident:

The jury, by its verdict, properly rejected the defense based on suicide and found that the death of the insured was due to accident. Indeed, it had no other course, for the proof as to suicide — when viewed in its most favorable aspect — did no more than raise an issue of fact, for it cannot be said to be so conclusive as to require the jury to return a verdict of suicide. When death has resulted from violence, the presumption against suicide does more than shift the burden of proof and upon having done so disappears from the case; it continues to the end of the case and if a fair question of fact is presented as to whether death was due to suicide or accident, then the jury should answer accident. (Emphasis added.)

Begley v. Prudential Insurance Company of America, 1 N.Y.2d 530, 154 N.Y.S.2d 866, 868, 136 N.E.2d 839, 841 (1956). The court simply ignored the language in Wellisch which had hinted that the burden of proof in an accident or double indemnity suit must always remain on plaintiff to prove fulfillment of the condition precedent to defendant's duty to pay even if that means that plaintiff must disprove suicide. It focused instead on the Wellisch policy which had disfavored judgments of "criminality or turpitude." Although the Begley court recognized that it was dealing with a double indemnity claim rather than a life insurance claim, its language was, if anything, more strict than Wellisch in its description of the burden to be imposed on a defendant insurer which claimed that its duty to pay was vitiated by the insured's suicide.

At a minimum it is clear from Wellisch and Begley that the presumption does more than shift the burden of producing evidence, but left unclear is exactly its incremental effect. One possibility is that the phrase in Begley "it continues to the end of the case" means the presumption itself constitutes "some evidence," sufficient as a matter of law to tip the scales in favor of the party claiming its benefit if a "fair question" is presented by the rest of the evidence. This construction would lead to the conclusion the court should direct a verdict for beneficiary if it makes a threshold determination a "fair question" exists. But this possibility is negated by the phrases in Begley "the jury should answer" and in Wellisch that it is "a rule or guide for the jury."

A second possibility is that the presumption shifts the ultimate burden of persuasion from the beneficiary to the insurance company, with the insurance company assuming the burden of convincing the jury of "suicide" by a preponderance of the evidence, just as it would in a straight life insurance policy excepting suicide. This would amount to giving the New York presumption the "Morgan" effect, p. 1081 supra. Such an interpretation gives meaning to the last quoted sentence in Begley: if a "fair question" of fact is presented, the jury "should answer accident" because the insurance company bears the burden and has not met it by a preponderance of the evidence. The difficulty with this possibility is the phrase in Begley that the jury had "no other course" once presented with a factual conflict than to find accident. This implies that the jury was somehow bound to reach its conclusion and that the presumption means something more than merely shifting the persuasion burden.

A third possibility, therefore, is that New York regards its presumption against suicide as analogous to a mandatory or "conclusive" presumption; i. e. if the underlying factual determination is made by the jury (the existence of a "fair question"), then the jury must find accident as a matter of law. This possibility gives a meaningful role to the jury, that of determining a "fair question of fact," and also gives effect to the strong language in Begley that the jury "had no other course," that it "should answer accident," and the description in Wellisch of the presumption as a "rule or guide for the jury."

Frankly, both possibilities two and three are fairly inferable from the language of Wellisch and Begley. Support for the view that the New York Court of Appeals now regards the Begley presumption as in reality a legal conclusion is found in the recent case People v. Leyva, 38 N.Y.2d 160, 379 N.Y.S.2d 30, 341 N.E.2d 546 (1975). The court, relying on McCormick's distinctions (p. 1081 supra) described the Begley "presumption" against suicide as a "rule of jurisprudence . . . which describe[s] a basic value judgment of the system. It is only such rules . . . which carry a requirement that the jury be told that, if it finds the facts in doubt, it must follow the ``presumption.'" 379 N.Y.S.2d 30, 37 note 3, 341 N.E.2d 546, 551 (first emphasis supplied). Leyva's use of "requirement" and "must," and its description of "rules of jurisprudence" leave little doubt that Begley is now regarded not as a true presumption in either the Thayer or Morgan senses but rather as a rule of law. Though Leyva was of course a criminal case, the discussion therein of Begley and the New York "presumption" against suicide is apparently applicable to any context, criminal or civil, in which the presumption may arise.

Following Wellisch's, Begley's and Leyva's instruction, we submitted the existence of a "fair question" to the jury. We did so in the following terms:

. . . a fair question is an honest question, one which is based on the evidence or lack of evidence in reason. A fair question is not, however, a vague, speculative, or fanciful question, or a possibility of a question, but is a real misgiving which results in there being a fair question in your mind.

Charge of the Court at 20. Of course, in doing so, we had already made the threshold determination that a jury could reasonably find the existence of a "fair question of fact," or else we would have granted defendant a directed verdict during trial. In our mind, a "fair question" by any definition of the term was present: nine days of expert testimony on aerodynamics and motive testimony about Jay Scott had supported both inferences of accident and suicide. We therefore made the threshold legal determination about which defendant complains in his instant motion as it addresses the propriety of the charge. In then submitting to the jury the existence or non-existence of a fair question of fact, we were asking the jury to perform exactly the function it performs in virtually every case — i. e. of evaluating the evidence against an everyday standard of common usage. In further instructing the jury that if they found a fair question to exist they "should find" accident we gave meaning to the strength of the presumption in Wellisch and Begley, to its description in Wellisch as a "rule or guide for the jury" and to Leyva's instruction that "the jury be told that, if it finds the facts in doubt it must follow the presumption."

We therefore conclude that defendant's contention that we improperly submitted a legal issue to the jury to be without merit. And we find insupportable defendant's further contention that the jury could not be instructed it must find for plaintiff if it found a fair question to exist.

VI. Sufficiency of Evidence to Support the Verdict

The defendant contends that it is entitled to a judgment n. o. v. (Fed.R.Civ.P. 50) or, at least, to a new trial on the ground that the verdict was against the weight of the evidence (Fed.R.Civ.P. 59). Those contentions need not detain us long.

The standard applicable to an n. o. v. motion was stated in Denneny v. Siegel, 407 F.2d 433, 439 (3d Cir. 1969), as being "whether, as a matter of law, the record is critically deficient of that minimum quantum of evidence from which a jury might reasonably afford relief." See 5A Moore's Federal Practice ¶ 50.07[2], at 50-76 (standards for granting judgment n. o. v. same as those for directing a verdict). Defendant contends that under the evidence no reasonable juror could have found the crash to have been an accident. Defendant insists that this is true even if New York's strong presumption against suicide is applied.

We disagree. Applying the Denneny standard to the evidence in this case, we would sustain a jury verdict for the plaintiff under either the New York "fair question" charge or under a traditional Pennsylvania charge that plaintiff must prove accident by a preponderance of the evidence. We view the evidence as creating a classical jury question. We note specifically that plaintiff's theories of the crash, as advanced by her highly qualified experts who testified that suicide was an impossible explanation and who proffered an alternative hypothesis, were sufficient to justify submitting the case to the jury. Additionally, if we were correct in applying New York law, the propriety of submitting the case to the jury cannot seriously be questioned because of the strength of the New York presumption against suicide. In sum, this case was most appropriate for a jury determination: there were two conflicting explanations of an event, each of which was supported by competent testimony.

Failing judgment n. o. v., defendant nonetheless requests that we grant it a new trial on the ground that the verdict was against the weight of the evidence. Such a motion is generally regarded as an appeal to the sound discretion of the trial court to prevent a jury from reaching a "seriously erroneous result." 6A Moore's Federal Practice ¶ 59.08[5], at 59-160. In granting such a motion a judge "does not act as a thirteenth juror in approving or disapproving the verdict." Id. In fact if a judge uses such a motion as an invitation to evaluate the credibility of witnesses and to resolve sharply conflicting testimony, he in turn invites reversal. Lewin v. Metropolitan Life Insurance Co., 394 F.2d 608, 615 (3d Cir. 1968). Again we believe that our summary of the evidence makes it clear that defendant's motion is precisely an invitation to be a thirteenth juror. For example, the testimony of Dr. Hubin, a physics professor as well as teacher of aerobatics and a most impressive witness, that a non-pilot's suicidal mania could not possibly have led to the maneuvers which were reported could, if believed, alone support this jury's verdict. Therefore, we decline the defendant's invitation. The motion for judgment n. o. v. or alternatively for a new trial on weight of evidence grounds must be denied.

VII. Admissibility of the FAA Airworthiness Directives

Defendant's final ground for seeking a new trial is our ruling admitting into evidence two FAA Airworthiness Directives (A.D.'s)^[69] (P-25 and P-26) which impugned the mechanical safety of that model of Piper Cherokee Arrow of which the plane that crashed was a specimen. Defendant's objection to the admission of the A.D.'s may fairly be characterized as contending that the A.D.'s were inadmissible as irrelevant, Fed.R.Evid. 401-403, as hearsay, Fed.R. Evid. 801-802, as opinion evidence of persons whose qualifications could not be properly tested by the Court, Fed.R.Evid. 702, and as opinion evidence the factual foundation for which could not be subjected to cross-examination at trial, Fed.R.Evid. 705.^[70] Furthermore, defendant argues that the admission of the Airworthiness Directives was prejudicial because one of plaintiff's expert witnesses, Professor Hubin, had relied on the A.D.'s to buttress his theory that the observed aerobatics resulted when the pilot tried, but failed, to compensate for a mechanical malfunction. Worse yet, the defendant suggests, is the fact that admission of the A.D.'s may well have appeared to the jury to have put a government imprimatur on plaintiff's theories. After reflection, we still believe that our ruling was correct.

We initially address defendant's relevancy objection, which rests on two pillars. First, defendant argues that because the A.D.'s were issued after the crash, they did not apply to the crashed plane. Clearly, to the extent that an A.D. is an order to perform certain maintenance, it would not apply to a plane that no longer existed at the time of its issuance. However, the factual finding (or judgment) underlying an A.D. — that a condition exists which may be dangerous — applies to each plane within the class described by the A.D. Thus, although in a negligence suit an A.D. issued after a crash would certainly be irrelevant to proving that the charter company had notice of the dangerous conditions, in a products liability action the same A.D. might be relevant to proving that a certain defect existed. In the instant case, where plaintiff in rebuttal needed to prove only that there were plausible alternative explanations of the crash, the post-crash discovery of a defect affecting the design or manufacture of the crashed plane was relevant beyond per-adventure.

Secondly, defendant insists that even if an A.D. might be relevant in the sense we just suggested, the law holds such putative evidence to be irrelevant on the issue of causation. Specifically, defendant recites the rule that "reports of other accidents" are excludable on the causation issue when similar circumstances have not been proved. Prashker v. Beech Aircraft Corp., 258 F.2d 602, 608-609 (3d Cir. 1958). While we believe that Prashker is sound, it is inapposite here. In Prashker there were significant variations in design among the aircraft whose crashes plaintiff sought to compare; additionally, because plaintiff there had the task of proving causation, the Court found reports of other crashes irrelevant in the absence of proof of similar circumstances at the time of each crash. In the instant case, the definition of the class to which a particular A.D. applies excludes planes whose designs are relevantly different from the planes in which the underlying problem was discovered. Also, under our choice of law ruling the plaintiff did not have the responsibility of proving the cause of the crash by a preponderance of the evidence, but had only to convince the jury that a non-suicide explanation was reasonable — or that there was a fair question whether the death was a suicide. Defendant's relevancy objection must therefore fail.

We treat together defendant's hearsay and opinion evidence objections. Both essentially turn on construction of Fed.R.Evid. 803(8), which provides that:

[r]ecords, reports, statements, or data compilations, in any form, of public offices or agencies, setting forth . . . (B) matters observed pursuant to duty imposed by law as to which matters there was a duty to report . . . or (C) . . . factual findings resulting from an investigation made pursuant to authority granted by law, unless the sources of information or other circumstances indicate lack of trustworthiness are admissible as exceptions to the hearsay rule. The conceptual underpinning for this exception, similar to that upon which other hearsay exceptions are based, see note 70 supra, is that public records or reports, by virtue of their being based on legal duty and authority, contain sufficient circumstantial guarantees of trustworthiness to justify their use at trial. Of course, as the last clause of Rule 803(8) makes clear, such presumptions of reliability may, in appropriate cases, be rebutted. See generally Weinstein, Evidence, ¶ 803(3)[01] (1976). Applying Rule 803(8) to the facts of this case, we observe that A.D.'s are promulgated pursuant to the FAA's statutory duty to prescribe "[r]easonable rules and regulations . . . governing . . . the inspection, servicing, and overhaul of aircraft . . . ." 49 U.S.C. § 1421(a)(3). We believe that A.D.'s are "factual findings resulting from an investigation made pursuant to authority granted by law," and that A.D.'s are therefore within the exception to the hearsay rule announced in § 803(8)(C). See also Berkebile v. Brantley Helicopter Corp., 462 Pa. 83, 337 A.2d 893 (1975) (A.D.'s fall within state law exceptions to hearsay rule). Although it is arguable that the proper ground of admissibility is § 803(8)(B), the Advisory Committee notes strongly suggest that "evaluative reports," to which genre the A.D.'s belong, come under the umbrella of § 803(8)(C) rather than 803(8)(B).^[71]

In light of defendant's focus on the A.D.'s as constituting opinion evidence, however, defendant might be deemed to suggest that the A.D.'s are not within the ambit of § 803(8) at all, and that the reports in fact contained opinion evidence for which a proper foundation was not established and which was presented by experts whose qualifications were not tested. This suggestion requires a further scrutiny of § 803(8), coupled with a consideration of the proper relationship between the official documents exception to the hearsay rule, Rule 803(8), and Rules 702 and 705, relating to the testing and qualification of expert witnesses and of the factual foundations on which they base their opinions. Defendant would insist that we should construe that relationship in accordance with the intention of the House Judiciary Committee

that the phrase "factual findings" [in 803(8)(C)] be strictly construed and that evaluations or opinion contained in public reports shall not be admissible under this Rule.

Notes of the Committee on the Judiciary, H.R.Rep. 93-650, 93d Cong., 1st Sess. 14, reprinted in [1974] U.S.Code Cong. & Admin.News pp. 7075, 7088. Defendant's argument would be that because of this House Report any non-factual element of the A.D.'s had to be excised from the texts before presenting them to the jury.

We believe that such a position would be incorrect, for it overlooks a significant part of the legislative history. The Senate Judiciary Committee was in sharp disagreement with its counterpart in the House on the interpretation to be given "factual findings" in 803(8)(C).

The House Judiciary Committee report contained a statement of intent that "the phrase ``factual findings' in subdivision (C) be strictly construed and that evaluations or opinions contained in public reports shall not be admissible under this rule." The committee takes strong exception to this limiting understanding of the application of the rule. We do not think it reflects an understanding of the intended operation of the rule as explained in the Advisory Committee notes to this subsection.

Notes of the Committee on the Judiciary, S.Rep. 93-1277, 93d Cong., 2d Sess. 18, reprinted in [1974] U.S.Code Cong. & Admin. News pp. 7051, 7064-65 [hereinafter Senate Notes]. This conflict between House and Senate leaves the expressed congressional intent at a standoff because no adjustment as to this issue was made in the language of the rule in conference. See Conf.Comm. Notes, H.R.Rep. 93-1597, 93d Cong., 2d Sess., reprinted in [1974] U.S.Code Cong. & Admin.News 7098.

As noted by the Senate Judiciary Committee, the Advisory Committee clearly intended that so-called "evaluative reports" would be included under 803(8)(C)'s "factual findings" unless, as the rule states, "sources of information or other circumstances indicate lack of trustworthiness." Thus, the Advisory Committee instructed that "the rule . . . assumes admissibility [of such evaluative reports] in the first instance but with ample provision for escape if sufficient negative facts are present." Factors that the Advisory Committee considered to be helpful in resolving doubtful cases included: (1) the timeliness of the investigation; (2) the special skill or experience of the investigating officials; (3) whether hearings were held and at what level; and (4) possible motivation problems underlying the investigation. The Senate Committee expressly endorsed this explanation of factors by the Advisory Committee, and concluded that "the language of the rule together with the explanation provided by the Advisory Committee furnish sufficient guidance on the admissibility of evaluative reports." Senate Notes, supra. The Advisory Committee's approach has similarly commended itself to commentators who have reviewed the rule. See, e. g., Weinstein, supra, ¶ 803(8)[03]; McCormick, Handbook of the Law of Evidence § 317 (2d ed. 1972). We agree with the Senate Judiciary Committee that the Advisory Committee's methodology provides adequate guidance and safeguards and are well satisfied that the A.D.'s were admissible under the aegis of Rule 803(8)(C).^[72]

Our conclusion is bolstered rather than undermined by the one case cited to us on this point by defendant, Atlantic Mutual Insurance Co. v. Lavino Shipping Co., 441 F.2d 473, 475 (3d Cir. 1971), a case which arose before enactment of the Federal Rules of Evidence. There the Third Circuit upheld the trial court's exclusion of a fire marshal's report, not because it contained an opinion, but for the very same reason that the fire marshal's live testimony was excluded: "the proffered testimony was too speculative and conjectural . . . for an expert cannot base his conclusion upon facts which are not warranted by the record." Thus, the report was excluded for just the sort of reason — circumstances indicating a lack of trustworthiness — that the Advisory Committee envisioned in drafting 803(8)(C). Compare Atlantic Mutual Ins. Co., supra, with Moran v. Pittsburgh Des Moines Steel Co., 183 F.2d 467 (3d Cir. 1950) (report of Bureau of Mines as to cause of gas tank explosion held admissible under then applicable business records exception).

Our reading of 803(8)(C) also harmonizes the rule with Rule 702.^[73] An initial finding that A.D.'s were admissible over a hearsay objection did not preclude defendant from raising an objection to the expertise of the F.A.A. Such a classic challenge to expert testimony is at the same time germane to the admissibility of the report under 803(8)(C) under the interpretation we have placed on that rule because the challenges go to "circumstances [that] indicate [a] lack of trustworthiness." Thus, under both 703 and 803(8)(C) the proper focus is on the qualification to express an opinion, not simply on the fact of expression itself. How we would have handled objections to the expertise of the F.A.A. we need not pause to speculate, because the objection was not so framed at trial. In the absence of such a specific challenge, defendant's position can only be that a court can never accept as competent expert evidence official reports that contain facts admixed with scientific or technical judgment unless the declarant is present. Such a reading of Rule 702 would emasculate Rule 803(8). This hardly seems wise if the policy judgment behind 803(8) is that, absent factors indicating untrustworthiness, such reports are both necessary and trustworthy sources of evidence. Weinstein, supra, ¶ 803(8)[01].

As to the interplay of Rules 803(8)(C) and Rule 705,^[74] similar harmonizing considerations apply. In the context of official evaluative reports, the presumption is in favor of the reports' admissibility in the face of a hearsay challenge. However, where the foundation of an opinion would be discredited under Rule 705 cross-examination, the presumptive trustworthiness of the opinion might well be sufficiently impugned to disqualify the report under 803(8)(C), thus requiring the proponent of the report either to produce the declarant for purposes of cross-examination or to forego use of the evidence. Once again, we observe that the foci of the hearsay rule and the opinion evidence rule, this time 803(8)(C) and 705, are the same: not the expression of an opinion, but rather, the foundation for it. In this case, defendant did challenge the foundation of the A.D.'s at trial but, after inquiry, we were satisfied that the defendant could point to no "circumstances indicat[ing] lack of trustworthiness" sufficient to rebut 803(8)(C)'s presumption of admissibility as to the A.D.'s that were in fact introduced into evidence.^[75] We remain so satisfied. As with Rule 702, we are not disposed to read Rule 705, which otherwise would give defendant the right to cross-examine plaintiff's expert as to the foundation of its opinion, so as to drain the vitality of Rule 803(8). To grant defendant's motion on the record of this case would in effect be to hold that although circumstances do not in fact indicate a lack of trustworthiness, official reports properly admissible under 803(8) become inadmissible merely because there is no opportunity to cross-examine the author of the report. We reject such a proposition.

In sum, we deny defendants' objections to the admission of the A.D.'s. Insofar as the reports were admissible, defendant's arguments based on prejudice are without force.

In consideration of the foregoing opinion we have this day entered an order denying defendants' motion for judgment n. o. v. or for a new trial.

NOTES

[1] Defendant has moved for judgment n. o. v. or alternatively for a new trial.

[2] We also believe that where both interest analysis and Restatement II are incapable of resolving the choice, the Pennsylvania courts would in all probability turn to the traditional rules of Restatement I, though the result we reach does not depend on such conclusion. See footnote 68 infra.

[3] Indeed, the evidence would have sustained a verdict under a charge more favorable to the insurer than the one we gave. That in no way conflicts with our judgment that if the New York charge was given in error, that error cannot be held to have been harmless.

[4] Final Pre-trial Order, ¶¶ 2, 3, 4, 6, 9-14.

[5] Ms. Barry married after this suit began and is now Virginia Barry Melville.

[6] When Ms. Barry first met Jay Scott in 1958, she was but a teenager, working as an unskilled "board girl" for Laird & Co., a stock brokerage firm for which Scott also worked. That same year she came under his supervision. When he left Laird & Co. to manage the Scott family holdings in 1961, she became his employee, although she had no office skills. During that period and until Scott's death, Ms. Barry served a dual role of mistress and secretary. Ms. Barry also bore Scott's child. She was paid only $500 a month; however, Scott paid all of her bills and living expenses; and made her substantial gifts of jewelry, cash, country club membership, vacations and a Mercedes-Benz, et alia. He even supported Ms. Barry's relatives. In addition it appears that Scott used Ms. Barry's checking accounts to hold large sums of money for his own purposes and that he convinced her to sign both his mother's and his wife's names on various documents. See Testimony of Virginia Barry Melville, N.T. at 103-124, 143-184.

[7] Testimony of L. L. Kirk at 6; testimony of A. E. Meihofer at 12-16.

[8] Testimony of L. L. Kirk at 8.

[9] Testimony of A. E. Meihofer at 15-24. "Porpoising" is the term used to describe "pitching up and down" like a porpoise. Id., at 15.

[10] Testimony of L. L. Kirk at 28.

[11] Testimony of G. H. Rhodes at 77-78.

[12] Plaintiff offered no proof of flaw in Halamka's extensive and unblemished record as a pilot except for a "technical" failure to have an updated health certificate.

[13] N.T. at 372-379 and 545-553.

[14] N.T. at 398, 563. It was the practice of the charter service to use only one pilot on its single-engine flights. On those flights a passenger could sit up front, where he would have access to the dual control yokes.

[15] N.T. at 793.

[16] N.T. at 530-537.

[17] In questioning the decedent's mother Jane Scott, Counsel elicited testimony that Mrs. Scott had received what she took to be proceeds of dividend checks made payable to her, that could only have been cashed if Jay, or someone at his bidding, were signing her name. Her knowledge might be taken as a basis for arguing that Jay was in fact authorized to sign for his mother or that Jay might in good faith have believed himself to have been so authorized. See N.T. at 428-429.

[18] As we have already noted, how the "presumption" is phrased is critical to the outcome of this case; here we speak only in general terms.

[19] N.T. at 721.

[20] An "Airworthiness Directive" (AD) is an official document of the FAA issued to warn owners and mechanics of serious mechanical problems that have been found — or reported — on some planes and for which check ups or other remedies are required on planes of the same model. See p. 1110 infra.

[21] N.T. at 714.

[22] See discussion of New York law in §§ IV(C) and V infra.

[23] It is true, as plaintiff suggests, that under the facts of this case, if Jay Scott committed suicide he also committed murder, hence plaintiff makes the argument that our law would recognize a presumption against murder even where "murder" is only a tangential finding in a civil case. However, we have not pursued our legal analysis in this direction.

[24] The question of classification is generally considered to be a function of the law of the forum. See e. g. Restatement II § 7.

[25] Restatement I defines jurisdiction as "the power of a state to create interests which under the principles of the common law will be recognized as valid in other states." § 42. The method of exercising legislative jurisdiction is said to be "by adopting certain principles of law as part of the common law of the state, or by an act of a legislative body." § 59. Finally, the actual exercise of legislative jurisdiction occurs "through its [the state's] law which creates interests upon the happening of the specified acts or events." § 60.

[26] Pilot Life was the converse of our case. The forum state had a stronger presumption against suicide than did the state of contracting. The insured had died in the state of contracting, but his widow was a domiciliary of the forum at the time of suit.

[27] The Comment to these sections of Restatement II explains as follows:

b. Rationale. Most rules relating to which party has the burden of persuasion are concerned primarily with questions of trial administration. When rules of this sort are involved, the forum will apply its own local law. . . .

Different considerations prevail in those situations where the relevant rule of the state of the otherwise applicable law goes beyond questions of trial administration and is primarily designed to affect decision of a particular issue. In such situations, the rule of the state of the otherwise applicable law will be applied. On which side of the line a given rule belongs may present a difficult problem for decision. A rule of general application dealing with the burden of persuasion is almost certainly concerned primarily with trial administration. Such a rule of the state of the otherwise applicable law will not be applied by the forum. On the other hand, a rule which singles out a relatively narrow issue from the general norm and gives it peculiar treatment may have been designed primarily to affect decision of the particular issue. . . . (emphasis added)

§ 133 Comment b. The Comment goes on to note that a rule of the latter sort will usually be set forth in a statute; however, it need not always be.

[28] We discuss infra at p. 1097 the inappropriateness in Erie terms of our use of § 302 in our pretrial bench opinion to conclude the choice of law issue. That error does not affect the present limited use of § 302 for classification analysis.

[29] Pilot Life, Buhler, and Weber all treat the judgment on the substance or procedure question as one for the forum to determine. That is certainly the position of Restatement I, see § 584; but it must not be construed to mean that the forum has no obligation to consider the issue afresh in the choice of law context. Any indication in non-choice cases that a rule is procedural is distinguishable by context. It should be emphasized, however, that what we are doing here is to make a prediction — this time about how Pennsylvania would classify New York's presumption against suicide in a choice of law context.

[30] Complaint, exhibit A.

[31] Indeed, the only factor raising a question about the nature of the policy is the "exceptions" provision. Under a life insurance policy, this exception would be crucial; under an accident policy it would be superfluous. This exception is, however, not enough to convert an accident policy into a life insurance policy.

[32] Complaint ¶ 5.

[33] McCormick, Evidence 803 (2d ed. 1972).

[34] Id. at 821.

[35] Id. at 826.

[36] Id. at 804.

[37] Id. at 803, 804.

[38] Other presumptions in Pennsylvania law, apparently stronger than the one against suicide, can shift the burden of producing evidence but not the persuasion burden, which remains on the party on whom it was originally cast. See Allison v. Snelling & Snelling, Inc., 425 Pa. 519, 229 A.2d 861 (1967), Waters v. New Amsterdam Casualty Company, 393 Pa. 247, 144 A.2d 354 (1958); McCormick, supra, at 826 note 68. Thus Pennsylvania can be described in general as subscribing to the "Thayer" theory of presumptions.

[39] See discussion at pp. 1107-1110 infra.

[40] Conflicts law generally has been in a state of great flux, if not turmoil, during the past twenty years. In discussing Restatement I, interest analysis, and Restatement II, we do not imply that these three describe the existing universe. See, e. g., the "True Rules" approach of Professor Ehrenzweig, the "Better Law" approach of Professor Leflar, and the eclectic theory of Professor Cavers. However, with the exception of Justice Robert's dissent in Cipolla v. Shaposka, 439 Pa. 563, 267 A.2d 854 (1970), which followed Leflar, and the majority opinion citing Cavers, only the three theories discussed herein have impacted upon Pennsylvania law.

[41] A variant of the interest approach, that espoused by Professor Ehrenzweig, places great emphasis on the interest of the parties as opposed to interest of the sovereign.

[42] One simple example of the contrast between the vested rights and the interest approach may be helpful. McSwain v. McSwain, 420 Pa. 86, 215 A.2d 677 (1966) was an intramarital suit for the wrongful death of a child. The couple lived in Pennsylvania, and the accident had occurred in Colorado. Colorado law permitted tort suits between husbands and wives, while Pennsylvania did not. Under Restatement I theory the court would have merely ascertained where the "wrong" had occurred, which was Colorado, and then applied Colorado law. However, using interest analysis instead, the Pennsylvania Supreme Court inquired into the legislative purposes behind the laws at issue, and found that the Pennsylvania and Colorado Legislatures had made differing estimates of the value of marital harmony. The court then reasoned that Colorado would have no interest in an application of its law to this case, whereas Pennsylvania would have an interest. Therefore, Pennsylvania law was chosen to apply. This situation presents a "false conflict" under interest analysis (see text above).

[43] Comment i to § 6 of Restatement II posits the prevention of "forum-shopping" as being an important value expressed in principle (f). This value was of paramount importance in Restatement I, since in theory if every state applied the same "system" to choose applicable law, then no matter in which state plaintiff sued, the same state's law would be chosen to govern. Thus there would be no advantage to plaintiff in "shopping" for one forum as opposed to another. For further discussion of forum shopping, see pp. 1103-1104 infra.

[44] While Elston contains elements of interest analysis, it does not follow Currie's classical interest theory (see p. 1083 supra) because there is no reference to a (comparatively) greater New Jersey interest displacing a presumptively valid lex fori. In one of his last essays, Professor Currie modified his view that lex fori should always apply in a true conflict, advocating that the forum re-examine its interests and give them a "moderate and restrained" interpretation, which would sometimes lead to subordination to the stronger interest of a sister state. See B. Currie, The Disinterested Third State, 28 L. & Contemp. Probs. 757 (1963). Justice Roberts could have followed this reasoning to reach his result in Elston, but there is no indication that he did.

[45] The statement that there was "no choice of law problem" has contributed to the confusion in the cases. In Boase v. Lee Rubber & Tire Corp., 3 Cir., 437 F.2d 527, 529 note 6 (1970), Judge Aldisert read that statement as supporting his view that "the Pennsylvania Supreme Court ha[d] not [yet] applied its rule in a contract case." As indicated above, our reading of Hunter is that the Pennsylvania Supreme Court had applied "interest analysis" and Restatement II to reach the conclusion that there was "no choice . . . problem." Nonetheless, it also could be contended that it had not applied "interest analysis" to a "contract case," because it had used "interest analysis" to categorize the issue as one of jurisdiction over adoption and not as one of the effectiveness of a contract, though we question the viability of such a distinction.

[46] Hunter is strong precedent in that its analysis was stated in terms of the "interest" of Pennsylvania and because it over-ruled a trial court's application of lex loci contractus. It is oblique only in one way: that it never stated that interest analysis could be applied to a contracts case; instead it categorized the case as noncontractual. Yet, that obliqueness was understandably interpreted by one panel of the Third Circuit as leaving the "contracts" question open. See Boase, note 45 supra.

[47] Justice Roberts took no part in the Danz decision.

[48] Griffith, McSwain, Kuchinic, and Hunter were false conflict cases preceding Gillan, all of which would have counselled choosing Pennsylvania not New York law under these facts, but none of those cases were analyzed.

[49] One other Third Circuit case straddles between Restatement I and "center of gravity" or interest analysis, apparently having concluded that either approach would lead to the same result. Mixing Equipment Co. v. Philadelphia Gear, Inc., 436 F.2d 1308, 1312 (3d Cir. 1971).

[50] We note that Jamison, the most recent Circuit decision on this issue, applied traditional rules, apparently attributing no significance to Gillan. See p. 1089 supra.

[51] As this opinion reached its final drafts, we encountered the opinion of the Third Circuit in DeMateos, Administratrix v. Texaco Inc. and Texaco Panama, Inc. (Texpan), 562 F.2d 895 (1977). In that case the Court of Appeals affirmed a District Court dismissal of a damage suit against Texpan on the grounds that the Eastern District of Pennsylvania was an inconvenient forum for the litigation. The principal thrust of the Court's opinion is directed towards an analysis (and explication) of the seven factors set forth in Lauritzen v. Larsen, 345 U.S. 571, 73 S.Ct. 921, 97 L.Ed. 1254 (1953), with the conclusion that all of these factors pointed to the inapplicability of American law (as opposed to Panamanian law which did apply). In the course of the opinion the Court noted that under the Jones Act, 46 U.S.C. § 688, the Eastern District of Pennsylvania was a permissible venue. It then added:

However, in considering the propriety of a venue transfer pursuant to 28 U.S.C. § 1404(a) the Court has emphasized that such a transfer should not, despite its convenience, result in a change in the applicable law. Van Dusen v. Barrack, 376 U.S. 612, 626-43, [84 S.Ct. 805, 11 L.Ed.2d 945] (1964). That principle is no less applicable to a dismissal on forum non conveniens grounds. (Emphasis added.)

In view of the differing procedural and factual situations between DeMateos and the present case and the fact that there did not appear to be a difference in the non-applicability of American law regardless of forum, we do not construe the court as holding that Van Dusen means that a state forum non conveniens dismissal carries with it the consequences which the present plaintiff ascribes thereto; i. e., that we must apply New York instead of Pennsylvania choice of law principles here because of the forum non conveniens dismissal in New York and the reinstatement of the litigation here.

[52] It is not implausible that some state will adopt the Van Dusen rule as its own way of choosing law when a foreign state dismisses on forum non conveniens. Such a rule could arguably advance a state's sovereign interests by encouraging a plaintiff to use its courts. On the other hand, such a rule might be hard for a state to accept where its interest is involved. In any case, such possibilities are mere speculation here.

[53] For suggestions that Erie's policy of encouraging uniformity between federal and state courts despite the accident of diversity may not be consistently maintainable at the intersection of forum non conveniens and § 1404(a) transfer, see Currie, Change of Venue and the Conflict of Laws: A Retraction, 27 U.Chi.L.Rev. 341, 348 (1960) ("[t]he problem now appears to be insoluble, in any completely satisfactory way, while diversity exists."); and In Re Air Crash Disaster, 399 F.Supp. 1106, 1121, 1122 (D.Mass.1975) (suggesting that "[u]niformity of result . . . is precluded whenever a federal court transfers an action which the state court would have dismissed," but concluding "that the forum non conveniens rules of the transferor forum do not affect applicable law after transfer" under § 1404(a)). We read Judge Caffrey's excellent discussion as being generally in accord with our views.

[54] Defendant tended to treat interest analysis as identical to the "significant contacts" approach of Restatement II.

[55] We will deal with this point at some length below. See p. 1100 infra.

[56] The distinction between "contracting" and "performance" is the major distinction in the contract provision of Restatement I. The Restatement admits it will not always be possible to draw a "logical" distinction between the two, but assures that a "practical" line can always be drawn. See § 358 comment b. Two years after Griffith, the Pennsylvania Supreme Court added a further puzzling distinction, in Silvestri v. Slatowski, 423 Pa. 498, 224 A.2d 212 (1966). It announced:

It has long been the law in Pennsylvania that in matters relating to the enforcement of a contract or in matters concerning the remedy for breach of contract, the lex fori, or the law of the place where such remedies are pursued applies.

Silvestri cites as support three cases, from 1899, 1885, and 1845. Id. at 215. The first, Musser v. Stauffer, 192 Pa. 398, 43 A. 1018 (1899) does not support the proposition; rather, like Restatement I, it makes "enforcement" and "remedies" for a "breach" functions of the place of performance, which may or may not be the forum. The 1845 and 1885 cases do support Silvestri. Therefore Silvestri seems to have resurrected a doctrine nearly one hundred years old that actually predates Restatement I. The Silvestri distinction is apparently between "rights" and "remedies," and may be analogous to the Restatement I distinction between "substance" and "procedure." Just as the forum automatically applies its law to "procedural" questions under Restatement I, without ever invoking conflict of law principles because these principles apply only to "substantive" matters, so too the forum under Silvestri and the XXXX-XXXX cases would automatically apply its law to all "remedial" issues. As Silvestri itself explained, it was not "concerned" with the "application of a conflict of law principle." Id. 224 A.2d at 215. We conclude Silvestri does not impact upon the case before us for two reasons. First, by determining the Pennsylvania Supreme Court would regard the presumption against suicide as "substantive," not "procedural," see § IV(B) supra, we have implicitly determined that it would regard the presumption as involving a "right," not a "remedy," within the framework of Silvestri. Second, Silvestri decided a jurisdictional issue — whether a Pennsylvania equity court could enforce a separation agreement made in another state. Thus, by its own terms quoted above, it did not purport to be, nor was it in fact, a statement about Pennsylvania choice of law methodology.

[57] [The policy was accepted by defendant] "Pursuant to a telephone call on or about July 8, 1966, from Johnson & Higgins of Philadelphia, Inc. to defendant [in New York] . . . an oral binder was effected. Thereafter, the written application . . . dated July 18, 1966 was mailed from the office of Johnson & Higgins, of Pennsylvania, Inc. . . . The policy was mailed on or prior to August 9, 1966 from the office of defendant [in New York] to Johnson & Higgins of Pennsylvania, Inc." Defendant's answers to plaintiff's interrogatories numbers 15-17.

[58] We do not doubt that there are situations where an insurer is registered to do business in a state other than its home state, and where it does such substantial business in that state that the non-home state may be so concerned to assure a sufficient fund for the legitimate claims of its citizens that it proposes to limit dubious claims which the insurer's home state might protect; but there is no evidence that such is the case here.

[59] Plaintiff's interrogatories, nos. 1-12, all sought to elicit information about the relationship between the insurer, the insurer's policy form and rates, and the sovereigns which might have regulated same. Pointedly, plaintiff asked whether there was any difference in rates on identical policies issued to residents of New York, Delaware, and Pennsylvania, whether there were any differences in the terms of the policies utilized in these states, and whether the policies had been submitted for administrative approval in any of those states. The defendant objected to all these interrogatories as irrelevant and failed to answer them. No motion to compel answers was filed and the trial record is silent on these points.

[60] Plaintiff's query as to whether defendant "transacted" insurance in Delaware (or in Pennsylvania) was among those not answered. Id.

[61] Compare Babcock v. Jackson, 12 N.Y.2d 473, 240 N.Y.S.2d 743, 191 N.E.2d 279 (1963) and Tooker v. Lopez, 24 N.Y.2d 569, 301 N.Y. S.2d 519, 249 N.E.2d 394 (1969) with Neumeier v. Kuehner, 31 N.Y.2d 121, 131, 335 N.Y.S.2d 64, 286 N.E.2d 454 (1972).

Most interest analysis cases recognize as state interests sufficient to trigger a sovereign's claim that its law should apply to a multi-state transaction only the state's interest in protecting its own domiciliaries whether as parties to the particular suit or as bystanders or persons who will benefit from deterrence of certain conduct in the future. See Cipolla v. Shaposka, 439 Pa. 563, 267 A.2d 854 (1970). The theory would doubtless lead in different directions if states were generally perceived as having interests in controlling the conduct of their own domiciliaries even to the extent of imposing duties on their relationships to foreigners in foreign places that the foreign state would not impose in purely domestic cases. For theoretical discussion of possible rational bases for such altruism, see Rawls, A Theory of Justice at 183-192 (1971); Nagel, The Possibility of Altruism (1970). Alternatively, a different "interest" theory might find that all states are interested in having the better rule of law applied to all situations and will believe their own to be such a better rule unless convinced otherwise. We avoid such alternative and speculative bases for "interests." Finally, we should note that some cases have defined "interests" in terms of a state's "contacts" with a particular relationship: marriage, family, etc. General "interest" in regulating a relationship within a state's territory seems to partake of and be consistent with both Restatement II and the interest approach. But how the notion is used will doubtless vary with the underlying theory. See, e. g., Grant v. McAuliffe, 41 Cal.2d 859, 264 P.2d 944, 42 A.L.R.2d 1162 (1953) (domicile controls decedent's estates); Haumschild v. Continental Casualty Co., 7 Wis.2d 130, 95 N.W.2d 814 (1959) (marital domicile controls interspousal immunity as matter of family law); Berghammer v. Smith, 185 N.W.2d 226 (Iowa 1971) (marital domicile controls wife's rights to recover for loss of consortium).

[62] See Ormsby v. Chase, 290 U.S. 387, 54 S.Ct. 211, 78 L.Ed. 378 (1933) and Dalton v. McLean, 137 Me. 4, 14 A.2d 13 (1940), both of which were decided under the place of injury rule without any analysis of the state's interest.

[63] In Labree v. Major, 111 R.I. 657, 306 A.2d 808 (1973), the Supreme Court of Rhode Island applied its own law to a case in the Erwin mold, but for a different reason. A Massachusetts guest of a Rhode Island driver was injured in an accident in Massachusetts. Rhode Island imposed a higher standard of care or a greater duty on a host to a guest than did Massachusetts in its domestic cases. The court did not couch its answer in simple lex fori terms, but insisted that its

policy of allowing recovery by passengers for the ordinary negligence of their hosts is not limited to the protection of Rhode Island guests. The state has an interest in enforcing the standard of care of an automobile operator no matter where his guest resides.

Id. at 306 A.2d at 816 (emphasis supplied). But following this usage of the terms policy and interest, the converse is also true: namely, if Massachusetts has a policy of protecting drivers from ungrateful guests, it also has an interest in enforcing its standard of gratefulness. What is crucial to "interest" analysis is to distinguish when states have particular interests in applying their policies; the policies themselves are always couched in universal terms.

If the language quoted from Labree were taken seriously, it would only imply that the state of which each party is a domiciliary is always "interested." We would then be left to choose between them. If this approach were combined with Currie's forum preference, then any discussion of the content of the laws of the states and of the policies underlying them would be rendered a mere facade. So long as one party came from the forum, forum law would apply. To avoid this reductio ad absurdum, we believe that Labree should be understood as applying lex fori to an unprovided for case.

[64] The Third Circuit followed this methodology in applying Pennsylvania conflicts law in Suchomajcz v. Hummel Chemical Company, 524 F.2d 19 (3d Cir. 1975). In this tort action, Judge Rosenn first analyzed relevant state interests, citing Griffith. Discerning that the two jurisdictions had incompatible interests, he found Pennsylvania conflicts law would look to territorial factors, in the wake of Cipolla, and would find that Pennsylvania, the "place of the injury," would provide the applicable substantive law. 524 F.2d at 23.

[65] See Monihan v. Monihan, 438 Pa. 380, 386-87, 264 A.2d 653, 655-656 (1970) (concurring opinion); Commonwealth ex rel. Blank v. Rutledge, 234 Pa.Super. 339, 344, 339 A.2d 71 (1975).

[66] In view of the role which we have suggested that territorial concerns still have in choice of law, it is appropriate to review briefly the theoretical basis upon which the fact that a state is the locus of an event generally constitutes a sufficient basis for that state's asserting its "sovereign decisional authority," Aldens, Inc. v. Packel, 524 F.2d 38, 43 (3d Cir. 1975), or its "legislative jurisdiction." Restatement I, § 62. See note 25 supra.

The claim of a sovereign to control events within its prescribed territory has long been considered as essential to its sovereignty. Indeed, territoriality has often claimed to be the exclusive basis for law-making and law applying power. There has, of course, been a breakdown of such exclusivity in a variety of contexts. Thus the rule of Pennoyer v. Neff, 95 U.S. 714, 733, 24 L.Ed. 565 (1877) that a court can bind only persons who are within its territorial jurisdiction at the time suit is brought, has given way to the principle of International Shoe Co. v. Washington, 326 U.S. 310, 66 S.Ct. 154, 90 L.Ed. 95 (1945), that the contacts necessary to support judicial jurisdiction must be analyzed flexibly and that presence is not required. Moreover, the holding in Home Insurance Co. v. Dick, 281 U.S. 397, 50 S.Ct. 338, 74 L.Ed. 926 (1930) that Mexico had exclusive legislative jurisdiction vis-á-vis Texas over a contract made and to be performed in Mexico, where the only bases for Texas' claim to provide the rule for the case were that Dick was a citizen of Texas at the time the contract was made and that Texas was the forum, see also Hartford Accident & Indemnity Co. v. Delta & Pine Land Co., 292 U.S. 143, 54 S.Ct. 634, 78 L.Ed. 1178 (1945) and Order of United Commercial Travelers of America v. Wolfe, 331 U.S. 586, 67 S.Ct. 1355, 91 L.Ed. 1687 (1947), has certainly been modified by Clay v. Sun Insurance Office Ltd., 377 U.S. 179, 84 S.Ct. 1197, 12 L.Ed.2d 229 (1964), which upheld Florida's power to apply its substantive law to an insurance contract made by a foreign citizen in a foreign state where the insured had become a citizen of Florida, where Florida was the forum, and where the defendant was licensed to do business in Florida.

International Shoe, however, denied only the exclusivity of "presence" as the basis for judicial jurisdiction, and Clay denied only the exclusivity of "territoriality" as the basis for legislative jurisdiction. Neither questioned the sufficiency of the territorial contact as a jurisdictional basis, but only its necessity. And neither suggested that the mere fact that a sovereign provides a forum for the resolution of a dispute is a sufficient ground for that sovereign to apply its own law, to the exclusion of the law of that sovereign within whose territory the disputed events occurred, unless the forum is "interested" in the dispute. Cf. Erie R. R. Co. v. Tompkins, supra.

Indeed the Supreme Court has suggested that Dick and Delta & Pine have continued vitality, not as support for the exclusivity of territoriality as a basis for legislative jurisdiction, but as holding that

It [is] a denial of due process of law when a state of the Union attempts to draw into control of its law otherwise foreign controversies, on slight connections, because it is a forum state.

Lauritzen v. Larsen, 345 U.S. 571, 590-591, 73 S.Ct. 921, 932, 97 L.Ed. 1254 (1953). And see, to similar effect Henry v. Richardson Merrell, Inc., 508 F.2d 28, 39 note 25 (3d Cir. 1974), DeMateos, Admx. v. Texaco, Inc., 562 F.2d 895 (3d Cir. 1977).

In candor it must be noted that the Lauritzen court found lex loci contractus to be as "fortuitous" as lex fori, and viewed lex loci delictus to be "of limited application to shipboard torts, because of the varieties of legal authority over water she may navigate." But this questioning of the sufficiency of "territoriality" in the context of maritime torts only reflects the special maritime rule that "the law of the flag supersedes the territorial principle." This in turn is justified either on the fictional ground that a ship remains within the territory of its sovereign or on the familiar pragmatic ground "that there must be some law on shipboard, that it cannot change at every change of waters, and no experience shows a better rule than that of the state that owns her." The Court concluded that

[t]hese considerations . . . must prevail unless some heavy counterweight appears.

345 U.S. at 588, 591, 583, 585, 586, 73 S.Ct. at 930.

In sum, we believe that our legal tradition recognizes a state's "concern" (we would call it a state's interest had that term not taken on a narrower meaning in this context) with events on the sole ground that they have occurred within its territorial confines as legitimate. That concern is legitimated in part by the symbolic import of territory in human affairs, see Ardrey, The Territorial Imperative (1966), and in part by the probability that events within its territory will affect the "interests" of the state or its citizens. It is also legitimated (see text supra) by the fact that it facilitates application of a principle (anti-forum shopping) which imparts into an important aspect of legal relationships a salutary regularity resting on sound policy.

[67] "The [Griffith] court wisely avoided the pitfall of substituting one rigid rule for another, but did go on to rely upon a balance of . . factors," Scott v. Eastern Air Lines, Inc., 399 F.2d 14, 22 (3d Cir.) (en banc) cert. denied 393 U.S. 979, 89 S.Ct. 446, 21 L.Ed.2d 439 (1968).

[68] The foregoing analysis does however suggest that, in a situation where pure Griffith interest analysis was first applied but produced an unprovided-for case, and where Restatement II was then applied but yielded a choice of law standoff, the Pennsylvania courts would, in view of the strong territorial overtones of its hybrid Griffith/Restatement II approach, be likely to turn to the traditional rules.

[69] An A.D. is a document issued by the Federal Aviation Administration (F.A.A.) describing "unsafe condition[s]" which have been found in some products and which, in the expert opinion of the Administrator of the F.A.A., are "likely to exist or develop in other products of the same type design." 14 C.F.R. § 39.1 (1977).

[70] We have found it necessary to "characterize" defendant's contentions for analytical purposes because, except for the relevancy objection, neither at trial nor in post-trial motions did defendant clearly distinguish the bases upon which its objection rested. At trial, the objection seemed to be one sounding in hearsay, N.T. 692, 698, but with overtones of inability to cross-examine as to foundation, N.T. 698. In the post-trial motions, the emphasis seems to have shifted, with the primary focus trained on inability to cross-examine and the hearsay objection introduced almost subliminally. Defendant's Brief in Support of Alternative Motions for Judgment Notwithstanding the Verdict or for a New Trial, 62-64.

Obviously, the hearsay and inability to cross-examine objections are closely related, for one of the primary purposes for excluding hearsay evidence is the inability to cross-examine the declarant. Under limited and carefully controlled circumstances, however, some evidence that would otherwise be inadmissible as hearsay is expressly made admissible by the Federal Rules of Evidence because of the inherent reliability of such evidence. See Advisory Committee Notes to Rule 803 ("The present rule proceeds upon the theory that under appropriate circumstances a hearsay statement may possess circumstantial guarantees of trustworthiness sufficient to justify nonproduction of the declarant in person even though he may be available."). Thus, it becomes extremely important to distill from a generalized hearsay objection the objection's specific grounds and then to analyze the objection in the framework of the Federal Rule's hearsay exceptions.

[71] Unlike 803(8)(C), 803(8)(B) does not limit admissibility to "factual findings," which, in a very strict sense, expert opinions such as those contained in the A.D.'s are not. The A.D.'s are, appropriate to 803(8)(B), based on "matters observed pursuant to duty imposed by law as to which matters there was a duty to report." Moreover, an "investigation," a prerequisite to a report under 803(8)(C), is paradigmatically an investigation of a specific occurrence in the past, rather than the forward-looking and class-oriented studies of which A.D.'s are a type. An investigation is not by its own terms limited to examination of past events, however, and the wording of the rule gives no indication that the term is to be so limited. Additionally, one might argue that "matters observed," 803(8)(B), are no less factual than "factual findings," 803(8)(C), because one can only "observe" objective phenomena. Judge Weinstein, for example, would limit 803(8)(B) to "facts ascertained through personal observation." Weinstein, Evidence, ¶ 803(8)[02] (1976). We prefer to rely on 803(8)(C) because of the intimations of the Advisory Committee noted in the text and because 803(8)(B) lacks, whereas 803(8)(C) contains, the salutary provision authorizing exclusion of the evaluative material when "the sources of information or other circumstances indicate lack of trustworthiness." Indeed, this fact would seem to confirm our reading of the Advisory Committee notes. See text accompanying note 72 infra.

We add only that we are unpersuaded by any argument that the admissibility of the A.D.'s is governed by 803(8)(B) which might be constructed upon the House Judiciary Committee's expressed intent to read "factual findings" in 803(8)(C) narrowly. Such an argument proves too much, for were that intent to govern our ruling, we would not be inclined to admit the A.D.'s under either (B) or (C); it is unlikely that the House would object to admitting expert opinion evidence under (C), but would be agreeable to admitting the very same evidence under (B), even though the latter lacks the untrustworthiness safeguard of the former.

[72] Indeed, to read 803(8)(C) as the House Judiciary Committee suggested might well be inconsistent with the spirit behind the Rules as evidenced in 803(6), the Federal Rules' business records provision. Rule 803(6) provides for the admissibility of a "memorandum, report, or data compilation, in any form, of acts, events, conditions, opinions, or diagnoses . . . unless the source of information or the method or circumstances of preparation indicate a lack of trustworthiness" [emphasis supplied]. To be sure, 803(6) expressly admits opinion while 803(8)(C) provides only for factual findings. Conscious, however, of the exceedingly hazy and uncertain nature of the line between fact and opinion, see, e. g. McCormick, Handbook of the Law of Evidence § 11 (2d ed. 1972); Weinstein, supra note 66, 803(8)[03]; 7 Wigmore, Evidence § 919 (3d ed. 1940), we observe that the two rules' escape mechanisms for excluding unreliable evidence are remarkably similar (803(6): when "the source of information or the method or circumstances of preparation indicate a lack of trustworthiness"; 803(8)(C): when "sources of information or other circumstances indicate lack of trustworthiness"). Such similarity in language connotes to us similarity in rationale and provides further indication that opinion evidence, expressly admissible under 803(6), at the very least should not be, as defendant would have it, per se inadmissible under 803(8)(C).

[73] Rule 702 provides:

If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise.

[74] Rule 705 provides:

The expert may testify in terms of opinion or inference and give his reasons therefor without prior disclosure of the underlying facts or data, unless the court requires otherwise. The expert may in any event be required to disclose the underlying facts or data on cross-examination.

[75] We excluded one A.D. at trial precisely because the circumstances of its compilation indicated a lack of trustworthiness, i. e., apparent reliance on hearsay. N.T. 707.

We note in this regard the recent opinion of the Third Circuit in John McShain, Inc. v. Cessna Aircraft Co., 563 F.2d 632 (3d Cir. 1977). There the court upheld the trial court's exercise of discretion in excluding accident reports of the National Transportation Safety Board under 803(8)(C) because "reception of such reports into evidence would have involved . . an inquiry into the ``circumstances bearing on the trustworthiness' of the investigators' conclusions, Fed.R.Evid. 803(8)(C)." at 636.

We do not consider the exclusion of the official reports in McShain to inveigh against our ruling in this case. The Third Circuit did not, as defendant would have us do, assume inadmissibility merely because the reports incorporated professional judgments. Rather than concerning itself with the mere fact that the reports contained conclusions, the Third Circuit focused, as have we, on the underlying trustworthiness of those conclusions. The reports at issue in McShain contained conclusions based in part on the hearsay testimony of witnesses to accidents. On such a record, conditioning receipt of the reports on "an inquiry into ``the circumstances bearing on [their] trustworthiness'" was appropriate. We perceive our ruling as to the excluded A.D. to be in accord with McShain on this point.

Regarding those A.D.'s that were admitted, no circumstances independently indicated a lack of trustworthiness, nor did defendant convincingly suggest any such circumstances. Indeed, we elicited testimony at trial that before promulgation of A.D.'s, notice is given and hearings are held. N.T. 695. The existence of such hearings is one of the trustworthiness indicators noted by the Advisory Committee. See text preceding note 69 supra. We do not read McShain to require in all cases an affirmative showing of trustworthiness beyond the prima facie showing required by the rule itself that the report consists of "factual findings resulting from an investigation made pursuant to authority granted by law." The clear language of 803(8)(C) and the explication of the rule given by the Advisory Committee establish that official reports are presumptively admissible unless the circumstances indicate a lack of trustworthiness. The "necessary" inquiry in McShain was prompted by the existence of such factors. In the absence of similar indicators, McShain does not require a sua sponte trustworthiness inquiry more extensive than that on which we embarked.