In this diversity action1 Sidney Roginsky sought to recover compensatory and punitive damages for personal injuries, primarily cataracts, from taking at his home in Pennsylvania2 a drug, MER/29, developed by Richardson-Merrell Company for lowering blood cholesterol levels. Roginsky's was the first to be tried of some 75 similar cases now pending in the District Court for the Southern District of New York. Several hundred actions have been filed elsewhere, see Rheingold, Products Liability — The Ethical Drug Manufacturer's Liability, 18 Rutgers L.Rev. 947-48 n. 4 (1964), in at least three of which trial courts have rendered large judgments for the plaintiffs.3

Although other theories of liability for compensatory damages had been advanced in the complaint, plaintiff withdrew all except negligence and fraud upon the Food and Drug Administration (FDA). Defendant moved for a directed verdict on all claims for injury by cataract as unsupported by sufficient proof of causation and on the fraud and punitive damage claims as unsupported by the evidence; the motions were denied. The judge instructed the jury it must first determine the issue of causation; if it found for the plaintiff on that, it should then pass upon the other issues, which he explained in a charge to which defendant took no exception. He helpfully submitted six separate questions: (1) causation, (2) negligence, (3) fraud upon the FDA, (4) amount of compensatory damages, (5) liability for exemplary damages, and (6) the amount thereof. The jury gave affirmative answers to all the questions relating to liability and fixed compensatory damages at $17,500 and punitive damages at $100,000, which the judge later declined to eliminate or reduce, 254 F. Supp. 430 (1966). On appeal defendant contends that its motions for directed verdicts should have been granted; it argues also that evidence erroneously admitted, much of it in support of what it considers the unsubstantiated fraud count, could have prejudiced the jury's determination of the issues of negligence and of conduct warranting the award of punitive damages. It also raises other objections to the receipt of evidence and complains that the award of punitive damages, "unless restricted to fixed and measurable amounts," violates due process. We affirm the award of compensatory damages but find that the evidence was not sufficient to warrant submission of the punitive damage issue to the jury.

A summary of certain undisputed facts from this enormous record will be useful by way of introduction. MER/29 was developed in the late 1950's by The Wm. S. Merrell Company, a division of defendant, for the purpose of lowering blood cholesterol levels. At that time most physicians believed that a high level of cholesterol was a significant precursor of atherosclerosis, the leading single cause of death in the United States.

Before the drug was placed on the market there had been 246 experiments involving 3907 animals and it had been administered to over 2000 human patients under close clinical observation. Eighty per cent of these patients who had used the drug for 90 days or longer experienced a reduction of cholesterol levels averaging 20%. The only reported side effects were dermatitis of several different types, two reports of hair loss, some nausea and vomiting, one report of a drop in white blood cell count, two cases of vaginal bleeding or spotting, three cases of tearing or watering of the eyes, and one of blurred vision. In December 1959 many of the clinicians who had been administering the drug reported to a conference of eminent scientists and physicians held under defendant's auspices at Princeton, N. J.

In July 1959 defendant filed with the FDA a New Drug Application (NDA) for MER/29. The FDA cleared the drug for release in April 1960, subject to submission of approved labels. The label stated, under the heading "SIDE EF-facts," that "Isolated reports have been received of nausea, vomiting, temporary vaginal bleeding and dermatitis" and under the heading "CAUTION" a warning that "MER/29 has been shown to be entirely safe in the periods the drug has been studied, but long-term or lifetime effects are unknown. Periodic examination of patients on long-term MER/29 therapy is therefore necessary." Marketing began June 1. During the balance of 1960 over 100,000 persons used the drug, with no reports of cataracts.

In January 1961, Merck & Co., which had borrowed a sample of MER/29 and then synthesized its own supply, reported that test animals had developed cataracts. Defendant, after sending a team to Merck's laboratory, decided to make a further experiment on animals selected by it but did not reveal the Merck report to the FDA or the medical profession. On February 1 it learned that a user in Los Angeles, one Lee Anticouni, had developed a cataract; the facts were never run down for reasons later developed, and neither the FDA nor the profession was informed. Meanwhile additional thousands of patients were using the drug, with benefit to cholesterol levels and for the moment without important reported adverse effects.

On defendant's rerun of the Merck experiment, its dogs developed cataracts in October, 1961. At the same time the Mayo Clinic reported cataracts in two, later three, patients who were using MER/29. The combined effect of these two incidents was to cause the defendant, on October 18, 1961, to request the FDA's permission to issue a warning letter to all physicians, such permission then being required by law, 25 F.R. 12595 (1960), 21 C.F.R. § 130.9 (1963 ed.). Believing defendant's proposed letter to be too weak, the FDA withheld permission and insisted on a much stronger one which was approved on November 27 and mailed on December 1 to every doctor in the country and also to defendant's salesmen. With additional reports of cataracts coming in from the field during the early months of 1962, defendant was on the point of holding a high-level intra-company conference to decide upon withdrawal of MER/29 when, about April 1, the Mayo Clinic reported cataracts in a six year old boy who had been given high dosages to counteract a severe case of excessive cholesterol unusual in a child. Early in April defendant decided to withdraw MER/29 from the market; this was done on April 17, 1962.

Plaintiff, then aged 60, began using MER/29 in February 1961. In June he noticed scaling, rashes and falling hair which he reported to his physician. These conditions became aggravated despite treatment; around the year-end he noted disturbing eye symptoms and stopped taking the drug. In about six months the skin and hair conditions disappeared but the eye ailment, later diagnosed as cataracts, became somewhat worse; however, it has not become sufficiently serious for him to have them removed.

Defendant's broadest contention is that a verdict should have been directed in its favor insofar as plaintiff sought damages for cataracts because of lack of adequate proof that these were induced by MER/29.4 Plaintiff relied on the testimony of a single physician, the chief of ophthalmology at the well-known Guthrie Clinic in the Robert Packer Hospital at Sayre, Pa. The physician testified that in November 1963 and various later dates he had found slight cataracts in both eyes and that he was of the opinion with a reasonable amount of medical certainty that these were caused by the taking of MER/29; he relied on "medical literature" and "conversations" that many persons who developed skin and hair changes like Roginsky's after taking MER/29 also developed cataracts.

Defendant claims this afforded too weak a basis for an opinion in view of the doctor's admission on cross-examination and the testimony of defendant's two experts that plaintiff's cataracts were in the deep cortex and nucleus of the lens where senile cataracts are most likely to develop, rather than nearer the surface where toxic cataracts usually manifest themselves. To this defendant added the testimony of its experts that the slow development of Roginsky's cataracts, which had been largely arrested at the time of trial, pointed to their being senile rather than toxic, and that his skin ailment had not been of the type that had generally preceded MER/29 induced cataracts. However, plaintiff's medical witness denied such universal distinctions between the manifestation of senile and toxic cataracts and between plaintiff's form of dermatitis and others, and the cross-examination of defendant's experts was not without effect. With all this supplemented by the abundant evidence that symptoms generally like Roginsky's had often been followed by toxic cataracts, the issue of causation was rightly left to the jury, even though on our own appraisal we might well find defendant's experts to be more convincing.

Defendant's next complaint is that an erroneous refusal to direct a verdict for it on the fraud count prejudiced the jury's consideration of the issues of negligence and of liability for punitive damages. While our holding that the evidence was insufficient to warrant submission of the issue of punitive damages moots the latter branch of defendant's complaint, the former remains; indeed its force is increased since one of plaintiff's arguments, that much of the evidence complained of was properly submitted to the jury on the punitive damages issue, falls by the wayside.

Plaintiff's further rejoinder, that any vice in the admission of evidence on the fraud counts was cured by the special verdict wherein the jury found defendant liable for negligence as well as for fraud, is likewise unacceptable. Defendant's objection is not that the jury lacked sufficient evidence to find negligence but that it may have acted on irrelevant evidence, since much material admitted on the fraud count dealt with sins of omission and commission that had no relation to plaintiff's injuries save in the sense, legally insufficient on a charge of negligence, that if defendant had done otherwise the FDA might not have released the drug at all.

Defendant asserts, and plaintiff does not dispute, that mere violation of the Food, Drug and Cosmetic Act does not give rise to a private claim — a point which, in view of the agreement of the parties, we do not decide. See Developments in the Law, The Federal Food, Drug, and Cosmetic Act, 67 Harv.L.Rev. 632, 722 (1954). Plaintiff relies instead on the principle that a fraud action will lie not only when the defendant practices his deceit directly on the plaintiff but also when he submits known false data to an intermediary who may be reasonably expected to pass it on, the classical case being the dispatch of financial statements to a credit agency. See Tindle v. Birkett, 171 N.Y. 520, 64 N.E. 210 (1902). Defendant counters that this principle applies only when the false information is to be passed on more or less as such, and not when it is to be used simply for an overall evaluation by a government agency whether the person submitting the data may put a product on the market. It argues on this basis that plaintiff should have been confined to introducing evidence of misinformation or failure to disclose information that might have affected the label or other communication whose form was subject to FDA approval. It contends also there was insufficient proof of reliance by the FDA on such false or incomplete data as were submitted, so that even "but-for" causation was not established.

One's liability generally for fraud has been said to extend only "to the persons or class of persons whom he intends or has reason to expect to act or to refrain from action in reliance upon the misrepresentation" and who in fact so rely, Restatement (Second), Torts § 531 (Tent. Draft No. 10, 1964), and in the specific case of information required to be filed by statute, to all persons who have justifiably relied thereon. Id. § 536. This would encompass a fraudulent misrepresentation to or concealment from the FDA affecting the approved label or from the patient's doctor if he is regarded as the plaintiff's agent, see Wechsler v. Hoffman-LaRoche, Inc., 198 Misc. 540, 99 N.Y.S.2d 588 (Sup.Ct. Bronx Co. 1950) (Rabin, J.). On the other hand to regard the FDA as "agent" for every person in the United States, as plaintiff urges, would be to use words as a substitute for substance. If we were forced to decide, we would say that a plaintiff does not make out a case of fraud simply by showing that if the facts had been fully stated, the FDA might not have released the drug.5 However, this is an area where the law is growing, see Prosser, Torts 717-19 (1964), and the question would become largely academic in new drug cases if an implied liability for violation of the Food, Drug and Cosmetic Act were to be recognized. Given our limited capacity as predictors of New York law, we would prefer to avoid decision of such a weighty and difficult issue, with wide ramifications as to other products, if we permissibly can.

We believe avoidance is justified because the proof of defendant's negligence, much of which is reviewed in section V of this opinion, was such that the jury's finding on that issue could not have been significantly influenced by admission of evidence on the fraud count. It seems to have been rather clear in this case from the outset that the real fight was about punitive damages, with causation and fraud important for their collateral effects on that issue, and not about negligence; indeed, we strongly suspect that if plaintiff had been willing to accept only compensatory damages, the action would never have had to be tried. The jury showed itself willing to find not only simple negligence but a good deal more; while the fraud evidence may well have affected the latter finding, it scarcely could have had an important effect on the former. We are most reluctant to put this plaintiff to another trial to recover compensatory damages which the jury was entirely warranted in giving; we shall therefore allow the judgment for compensatory damages to stand.

We thus come to the issue of punitive damages, an issue of extreme significance not only in monetary terms to this defendant in view of the hundreds of pending MER/29 actions and to the plaintiff as well, but from a longer range, to the entire pharmaceutical industry and to all present and potential users of drugs. Plaintiff, of course, does not claim that defendant intended to harm him; his contention is that defendant's negligence rose to such a level of irresponsibility or worse as to invite this extraordinary sanction.

The remedy has a long history. Its first articulation in England came in a case of illegal entry. The jury was held justified in going beyond "the small injury done to the plaintiff" because of the desirability of taking account of "a most daring public attack made upon the liberty of the subject" through entry and imprisonment pursuant to "a nameless warrant." Huckle v. Money, 2 Will.K.B. 206, 95 Eng.Rep. 768 (1763). See also Wilkes v. Wood, Lofft 1, 18, 19, 98 Eng. Rep. 489, 498-99 (C.P.1763). Later decisions reflect a variety of rationales: redressing affronts to personal feelings not susceptible of measurement, cf. Tullidge v. Wade, 3 Wils.K.B. 18, 95 Eng. Rep. 909 (1769), financing the cost of deserving litigation where only small compensatory damages can be expected, diverting the plaintiff's desire for revenge into peaceful channels, and serving as punishment for and deterrence from socially disapproved conduct.6 "Typical of the torts for which such damages may be awarded are assault and battery, libel and slander, deceit, seduction, alienation of affections, malicious prosecution, and intentional interferences with property such as trespass, private nuisance, and conversion." Prosser, Torts § 2 at 10-11 (1964). What strikes one is not merely that these torts are intentional but that usually there is but a single victim; a punitive recovery by him ends the matter, except for such additional liability as may be provided by the criminal law.

There is no doubt, however, that the remedy has been extended to cases where, although the defendant did not intend to harm the plaintiff, he showed "such a conscious and deliberate disregard of the interests of others that his conduct may be called willful or wanton." Prosser, Torts § 2 at 10 (1964). Such an extension was altogether natural: from a moral standpoint there is not too much difference between the driver who heads his car into a plaintiff and the driver who takes the wheel knowing himself to be so drunk that he probably will hit someone and not caring whether he does or not; and it is as important to deter the latter type of conduct as the former. But such cases still resemble those first considered in an important respect — a high probability that the number of plaintiffs will be few and that they will join, or can be forced to join, in a single trial.7

The legal difficulties engendered by claims for punitive damages on the part of hundreds of plaintiffs are staggering. If all recovered punitive damages in the amount here awarded these would run into tens of millions, as contrasted with the maximum criminal penalty of "imprisonment for not more than three years, or a fine of not more than $10,000, or both such imprisonment and fine", 21 U.S.C. § 333(b), for each violation of the Food, Drug and Cosmetic Act with intent to defraud or mislead.8 We have the gravest difficulty in perceiving how claims for punitive damages in such a multiplicity of actions throughout the nation can be so administered as to avoid overkill. Judge Croake did all that he could here, instructing the jury that it "may consider the potentially wide effect of the actions of the corporation and, on the other hand, * * * the potential number of actions similar to this one to which that wide effect may render the defendant subject."9 Yet it is hard to see what even the most intelligent jury would do with this, being inherently unable to know what punitive damages, if any, other juries in other states may award other plaintiffs in actions yet untried. We know of no principle whereby the first punitive award exhausts all claims for punitive damages and would thus preclude future judgments;10 if there is, Toole's judgment in California, see fn. 3, which plaintiff's brief tells us came earlier, would bar Roginsky's. Neither does it seem either fair or practicable to limit punitive recoveries to an indeterminate number of first-comers, leaving it to some unascertained court to cry, "Hold, enough,"11 in the hope that others would follow. While jurisprudes might comprehend why Toole in California should walk off with $250,000 more than a compensatory recovery and Roginsky in the Southern District of New York and Mrs. Ostopowitz in Westchester County with $100,000, most laymen and some judges would have some difficulty in understanding why presumably equally worthy plaintiffs in the other 75 cases before Judge Croake or elsewhere in the country should get less or none. And, whatever the right result may be in strict theory, we think it somewhat unrealistic to expect a judge, say in New Mexico, to tell a jury that their fellow townsman should get very little by way of punitive damages because Toole in California and Roginsky and Mrs. Ostopowitz in New York had stripped that cupboard bare, even assuming the defendant would want such a charge, and still more unrealistic to expect that the jury would follow such an instruction or that, if they didn't, the judge would reduce the award below what had become the going rate. There is more to be said for drastic judicial control of the amount of punitive awards so as to keep the prospective total within some manageable bounds. This would require, for example, a reduction of the instant $100,000 award to something in the $5000-$10,000 range, still leaving defendant exposed to several million dollars of exemplary damages. We perceive nothing in the New York decisions that would prevent our reducing a punitive damage award because of the large number of suits arising out of the same conduct by the defendant.12 But there is equally nothing to indicate that New York would follow such a course, and a state otherwise willing to impose such self-denying limits might be disinclined to do so until assured that others would follow suit.

Although multiple punitive awards running into the hundreds may not add up to a denial of due process, nevertheless if we were sitting as the highest court of New York we would wish to consider very seriously whether awarding punitive damages with respect to the negligent — even highly negligent — manufacture and sale of a drug governed by federal food and drug requirements, especially in the light of the strengthening of these by the 1962 amendments, 76 Stat. 780 (1962), and the present vigorous attitude toward enforcement, would not do more harm than good. A manufacturer distributing a drug to many thousands of users under government regulation scarcely requires this additional measure for manifesting social disapproval and assuring deterrence. Criminal penalties and heavy compensatory damages, recoverable under some circumstances even without proof of negligence, should sufficiently meet these objectives, see Note, supra note 6, 41 N. Y.U.L.Rev. at 1171, and the other factors cited as justifying punitive awards are lacking.13 Many awards of compensatory damages doubtless contain something of a punitive element, and more would do so if a separate award for exemplary damages were eliminated. Even though products liability insurance blunts the deterrent effect of compensatory coverage under such policies is often awards to a considerable extent, the total limited, bad experience is usually reflected in future rates, and insurance affords no protection to the damage to reputation among physicians and pharmacists which an instance like the present must inevitably produce. On the other hand, the apparent impracticability of imposing an effective ceiling on punitive awards in hundreds of suits in different courts may result in an aggregate which, when piled on large compensatory damages, could reach catastrophic amounts. If liability policies can protect against this risk as several courts have held,14 the cost of providing this probably needless deterrence, not only to the few manufacturers from whom punitive damages for highly negligent conduct are sought but to the thousands from whom it never will be, is passed on to the consuming public; if they cannot, as is held by other courts and recommended by most commentators, a sufficiently egregious error as to one product can end the business life of a concern that has wrought much good in the past and might otherwise have continued to do so in the future, with many innocent stockholders suffering extinction of their investments for a single management sin.

However, the New York cases afford no basis for our predicting that the Court of Appeals would adopt a rule disallowing punitive damages in a case such as this, and the Erie doctrine wisely prevents our engaging in such extensive law-making on local tort liability, a subject which the people of New York have entrusted to their legislature and, within appropriate limits, to their own courts, not to us. Our task is the more modest one of assessing the sufficiency of the evidence within the framework of New York decisions on the award of punitive damages for recklessness.15 As to this, we are convinced that the consequences of imposing punitive damages in a case like the present are so serious that the New York Court of Appeals would subject the proof to particularly careful scrutiny.16

The parties are in substantial agreement on one point — that New York does not impose punitive damages on a corporation unless, as charged by Judge Croake, "the officers or directors, that is, the management" of the company or the relevant division "either authorized, participated in, consented to or, after discovery, ratified the conduct" giving rise to such damages. Cleghorn v. New York Cent. & H. R. R. R., 56 N.Y. 44, 47-48 (1874); Wright v. Glen Falls, S. H. & Ft. E. St. R. R., 24 A.D. 617, 48 N.Y.S. 1026 (3d Dept. 1898); Rowe v. Brooklyn Heights R. R., 71 A.D. 474, 75 N.Y.S. 893, 894 (2d Dept. 1902); Walker v. Lord & Taylor, 236 A.D. 111, 258 N.Y.S. 96 (1st Dept. 1932).17 New York, in other words, adheres to the "complicity rule," holding the corporate master liable for punitive damages "only when superior officers either order, participate in, or ratify outrageous conduct." See Morris, supra, note 6, 21 Ohio St.L.J. at 221.18

The New York courts have used a variety of phrases to describe the "moral culpability," Walker v. Sheldon, 10 N.Y.2d 401, 404-405, 223 N.Y.S.2d 488, 491, 179 N.E.2d 497 (1961), which will support punitive damages for nonintentional torts.19 In Caldwell v. New Jersey Steamboat Co., 47 N.Y. 282, 296 (1872), the Court of Appeals spoke of "utter recklessness." Later it said the conduct "must be reckless and of a criminal nature, and clearly established." Cleghorn v. New York Cent. & H. R. R. R., 56 N.Y. 44, 48 (1874). It has also used such terms as "wanton or malicious, or gross and outrageous," Powers v. Manhattan Ry., 120 N.Y. 178, 182, 24 N.E. 295, 296 (1890), and "conscious indifference to the effect of his acts." Gostkowski v. Roman Catholic Church of the Sacred Hearts, 262 N.Y. 320, 323, 186 N.E. 798, 799 (1933). Very recently the Third Department has spoken of action "`committed recklessly or wantonly, i. e., without regard to the rights of the plaintiff, or of people in general,'" Soucy v. Greyhound Corp., 27 A. D.2d 112, 113, 276 N.Y.S.2d 173, 175 (3d Dept. 1967), quoting Magagnos v. Brooklyn Heights R. R., 128 A.D. 182, 112 N.Y.S. 637, 638 (2d Dept. 1908). In an earlier decision that court had held that "culpable negligence" was not enough to warrant punitive damages. Noonan v. Luther, 119 A.D. 701, 104 N.Y.S. 684 (3d Dept. 1907). What comes through from a study of these and many other New York decisions is that the recklessness that will give rise to punitive damages must be close to criminality, see 14 N.Y.Jur., Damages § 181 p. 41, and that, like criminal conduct, it must be "clearly established."20 It therefore seems appropriate to look to the latest authoritative New York definition of recklessness, that expressed by the legislature in the Revised Penal Law, McKinney's Consol.Laws, c. 40, § 15.05, subd, 3:

"A person acts recklessly with respect to a result * * * when he is aware of and consciously disregards a substantial and unjustifiable risk that such result will occur or that such circumstance exists. The risk must be of such nature and degree that disregard thereof constitutes a gross deviation from the standard of conduct that a reasonable person would observe in the situation."

Obviously this definition would be met if a manufacturer placed a drug on the market without any test program, a practice now rendered unlawful by federal legislation, or when its management knew the program had disclosed dangers of serious mischance or was incomplete in some material respect. Sufficient proof would also be furnished if, after the drug had been placed on the market, the manufacturer was shown to have become aware of danger and to have done nothing, deliberately closing its eyes. On the other hand, error in failing to make what hindsight demonstrates to have been the proper response — even "gross" error — is not enough to warrant submission of punitive damages to the jury. Cf. De Marasse v. Wolf, 140 N.Y. S.2d 235 (Sup.Queens Co.1955).

The judge properly did not restrict the plaintiff to evidence showing that defendant should have been aware that MER/29 was cataractogenic, since this other evidence tended to show the defendant's overall attitude toward the drug including its danger to the eyes. See Voltz v. Blackmar, 64 N.Y. 440, 445 (1876). However, the presence of the fraud count led to the reception of a great deal of evidence that was not properly cognizable on the issue of punitive damages because it was insufficiently connected with the management. If the fraud count should have been dismissed, as we tend to think, submitting this evidence to the jury would call for a new trial as to punitive damages.21 But we need not decide this point, because of our holding that there was insufficient evidence of management misconduct submitted to the jury to justify a verdict of punitive damages. An attempt to analyze every item of evidence relied on by plaintiff would prolong this opinion to a length even more inordinate than it has assumed; we limit ourselves to the evidence deemed most significant and helpful to the plaintiff.

One of the earliest and most inflammatory items was this: As part of the New Drug Application submitted to the FDA, defendant reported on a 16 month study of three pairs of monkeys, one of each pair being given the drug and the other being a control. The plan was to start the dosage at two and a half times the anticipated human dose, and then to raise it first to five and then to ten times. The experiment had been conducted by a toxicologist, Smith, who had left the company's employ before reporting on it; the write-up by his successor, Dr. King, contained numerous errors — none, however, shown to have been known to management — and one thing that was much worse. A laboratory technician testified to an occasion when she had submitted a final graph of the weights of the monkeys to a Dr. Van Maanen who directed her to increase the weight shown for one monkey whose weight had dropped 25% and to show weights for others for two weeks after they had in fact been autopsied; after complaining to Dr. King, her immediate supervisor, she made the changes. But Dr. Van Maanen, Merrell's Director of Biological Science, was also a subordinate who reported to the vice president and research director, Dr. Werner, and there was not the slightest evidence that the latter or any higher authority knew of his dictating this change in the observed data.22 Equally irrelevant to the issue of punitive damages was another error by Dr. King in reporting an experiment on female rats, all of which were shown to have died under exceptionally heavy dosage. Here also there was no proof of management complicity, nor was there any showing that Dr. King's error was deliberate. Moreover, it has not been clearly shown that either of these misreports was truly material.

The next episode relates to a study of 90 rats ending February 1960, of which thirty received ten times the recommended human dose, thirty received five times, and thirty served as a control, with a third of each to be autopsied in three, six and twelve months respectively. At the first autopsy corneal opacities (not cataracts, which are lenticular opacities) were observed in eight out of the twenty drugged rats. This was known to Getman, president and general manager of the Merrell Division, and was reported to the FDA. Dr. Murray, assistant to the Vice President in charge of Research, wrote the FDA that a consultant had advised that such corneal changes were common in laboratory rats and he believed them to be inflammatory. The record affords no basis for believing this was not the company's sincere opinion. Later there was an increased incidence of cataracts in the drugged rats as well as in some of the control rats; the former was not reported although the latter was, but there is no evidence that management was aware of this failure.

Dr. King also failed to report either to management or the FDA the development of cataracts in two dogs on a three month study ending February 1960, as to which he also modified his description of other abnormalities. Again there is no evidence that management was aware of this; furthermore, the whole study was worthless since the dogs were contaminated by distemper and viral hepatitis, and when it was rerun with sound dogs, no opacities were observed.

Plaintiff makes much of a memorandum dated May 17, 1960, shortly before MER/29 was put on the market, between two company vice-presidents with copies to still higher officers, concerning an analog which had been developed in the course of experimentation. Seizing on a phrase which suggested the desirability of having the analog "available as a substitute in the event Mer-29 gets into trouble," plaintiff argues this supports an inference that defendant put MER/29 on the market knowing that it would get into trouble, at the very time when a better product was in hand. But plaintiff takes the phrase quite out of context. The fact is that the letter discusses the company's general policy on licensing its drugs to other manufacturers and the application of this policy to the MER/29 analog. It says that animal work performed to date "indicates some question as to the toxicity of the analog in the lower animals (just as MER-29 itself does), but the work with the analog indicates it may be even better than MER-29. Although a lot more animal work would have to be done to tie down better the question of toxicity, action of the product, etc., this could be done. * * * In view of the tremendous importance of MER-29 and this analog, it probably will be decided that Merrell should follow through with the additional animal work and the clinical work on this analog in order to prove out its relative inferiority or superiority and thus have it available as a substitute in the event MER-29 gets into trouble, or perhaps introduced later as an improved product." The memorandum goes on to discuss various pros and cons of licensing an analog not thoroughly tested to another pharmaceutical manufacturer who would do the work. We find no basis for the sinister inference plaintiff would draw; the picture is rather of a company reasonably satisfied that it had a good product despite some known "toxicity in the lower animals" and considering what to do with another that might or might not prove better.

This brings us to the receipt of the report from Merck & Co. in January 1961 as to cataracts in dogs which we mentioned at the outset. Since this, to our minds, was the first information reaching management which might indicate that MER/29 could have such serious side effects as to make it unsafe despite its beneficial qualities, the officers' action must be severely scrutinized.

Defendant's immediate response was to send Dr. Werner, its research vice-president, Dr. Van Maanen and Dr. King to the Merck laboratory. The decision made on their return was to rerun the experiment. On its face this was not heedless or reckless conduct;23 on the contrary, to have accepted the Merck report as conclusive would have been irresponsible. Apart from the possibility of chemical differences in the Merck synthesized drug, a point on which defendant heavily relies but to which plaintiff effectively responds that Merrell did not ask for a sample,24 significance of the Merck results was called in question by the dogs' being beagles, which are known to be cataract prone, and defendant's suspicion of inbreeding which would have further invalidated the results.

Indeed, the serious criticism is not over what defendant did but what it failed to do — notify the FDA, the medical profession, or both. Plaintiff does not contend that FDA regulations required defendant to report the Merck results, the requirements then being limited to reporting results of experiments done in the manufacturer's own laboratory or by an investigator who had been furnished the drug. See 21 C.F.R. § 130.4(c) (1963 ed.). The claim is rather that, even though there is no direct proof that such notification was considered and rejected, its logic was so apparent as to permit an inference that defendant didn't want the FDA to know of the Merck results for fear it might order the warning accompanying the drug to be stiffened or the drug withdrawn, and didn't want physicians to know for fear they would be unduly frightened. Assuming a jury could find that this was so and that defendant was motivated by commercial considerations as well as by a desire to benefit potential victims of atherosclerosis and granting that management knew its own experiments had disclosed corneal opacities in rats, we fail to see how this meets the definition of reckless indifference to human life or health. The question on that score is not what defendant feared the FDA or doctors would do but what it feared MER/29 would do. Here the Merck report, with the infirmities we have noted, had to be set against the belief of defendant's management, erroneous only in one minor respect, that no cataracts had ever been reported in the 4000 animals previously tested including monkeys, which are closer than dogs to humans, in the 2000 patients under clinical observation for many months, or in the 100,000 patients who had used the drug for varying periods since its release in June 1960. We cannot see how these circumstances would warrant allowing a jury to find that Getman and Werner, the two principal decision-makers on this matter, who continued to use MER/29 themselves, consciously disregarded a substantial and unjustifiable risk of which they were aware.

Neither do we think the scale is tipped by the Anticouni incident, which followed shortly thereafter. Dr. McMaster, associate director of medical research for the Merrell Division, had first heard of Anticouni in May 1960 from a Los Angles cardiologist, one of the clinicians who had been administering the drug on an experimental basis. The cardiologist then reported that Anticouni had exhibited very severe dermatitis which he was unable to evaluate due to the patient's taking, in addition to MER/29 at twice the recommended dose, several other drugs, one of which has since been removed from the market because of its production of toxic side effects including severe dermatitis. McMaster asked for the records and offered to provide funds for further investigation; on several later occasions he repeated the former request but without result. On February 1, 1961, the cardiologist called Dr. McMaster to report that Anticouni had left his care and was said to have developed cataracts and to be consulting other physicians and attorneys. McMaster again asked for the cardiologist's files but did not receive them. He did nothing more until June when, on a trip to Los Angeles, he examined the files but was not enlightened, and sought to obtain the name of Anticouni's ophthalmologist without success. This episode was known to Getman, at least in a general way. Despite defendant's argument that the report was mere hearsay, that the other drugs Anticouni was taking were as likely as MER/29 to have caused cataracts, and that a cataract in this cardiac patient was likely to be of the senile, not the toxic, variety, a jury could find that the failure to exert greater energy in running down this report was negligent.25 However, McMaster's sloth would not support a finding of recklessness unless one assumes the evidence warranted a finding that by this time defendant had reason to be seriously apprehensive that MER/29 was cataractogenic in humans, an assumption for which we find no sufficient basis.

Before continuing with developments as to cataracts, we must recount another episode on which plaintiff heavily relies because the participation of management is plain. On March 24, 1961, Vice President Woodward of the Merrell Division addressed a memorandum to President Richardson of the parent company, with copies to Getman, Werner, McMaster and Murray, as to side effects of MER/29 on hair, a phenomenon which later experience indicated to be a not infrequent precursor of cataracts. Woodward reported that the company "had initiated a review of reported thinning of the hair of people taking MER/29," that this had disclosed 51 such reports out of 300,000 users, and that he and Getman felt "morally and legally bound" promptly to alter the warning accordingly. He recommended giving this warning despite the risk that it might provide the FDA with an occasion to re-examine Merrell's entire NDA: "we have made no changes to this point in any of our MER/29 literature, basically because we were afraid to `stir the pot' in Washington" since "we have heard from several sources that FDA at times has considered reopening our NDA file but, frankly, we do not know whether this is true." He enclosed a draft of a proposed letter to the FDA referring "to several reports concerning thinning or texture changes of hair" and proposing to add to the warning in the brochure accompanying MER/29 the words "and thinning of the hair." Richardson referred the memorandum to another vice president of the top company, Stormont, who promptly reported that although the facts as to hair effects were inconclusive, "the weight of the evidence seems to indicate that MER-29 may produce some hair changes in a very small proportion of cases" and that the warning should be revised as proposed. So far fine, but Stormont added that he didn't "particularly like" Dr. McMaster's suggestion that the warning be broadened to read "changes in color, texture, or amount" instead of "loss," since "this sounds rather frightening"; McMaster's memorandum had disclosed that of the 51 cases of hair changes, only five had included color and six texture. Dr. McMaster agreed with Stormont's suggestion. Richardson authorized proceeding along these lines,26 and the letter, limited to thinning, went forward on April 6. To us this seems much ado about rather little. The company recognized its obligation to inform the FDA and the public of what it had learned about hair thinning even though this had been reported in only 0.017% of the users. Decision not to advise of the other hair effects reported by 0.003% of the users does not fit our notion of reckless and wanton conduct.

Since the Merck report had included rats27 as well as dogs, defendant had initiated a rat study in addition to that on dogs. The rats began to develop eye difficulties in June and August, but there is no evidence that management learned of this until much later, when the experiment was nearing completion. What they did learn — on October 17, 1961 — was that although no ocular changes in the dogs had occurred as late as September 21 (including an examination on September 20 in which a Merck scientist had participated), a dog had developed a cataract by September 28, exactly six months after the onset of MER/29 therapy, and had become blind a week later;28 the report also told of the development of corneal opacities in most of the rats. This bad news synchronized with information from the Mayo Clinic at Rochester, Minnesota, that two patients taking MER/29 had been found to have cataracts "and that there was a question of the cataracts having been associated with administration of MER/29." Dr. McMaster went immediately to Rochester, learned of a third patient who had developed a cataract, returned to New York, and recommended that all physicians be sent a warning letter. It is common ground that, as FDA regulations then stood, 21 C.F.R. § 130.9 (1963 ed.), advance permission by the agency was required.

The proposed letter to the medical profession was first read to the FDA over the telephone on October 18. It began by saying its purpose "is to advise you of those cases where MER/29 therapy should be discontinued." It recited that there had been "a number of reports of hair loss, changes in color and texture of hair, and dermatitis" and that "more recently, lens opacities have been observed in four patients following severe dermatitis." It advised immediately withdrawing MER/29 "if changes in hair or skin occur" and warning patients to watch for and report such symptoms. The proposed letter went on to describe the types of dematitis and hair loss that would indicate cessation of therapy; it concluded by saying that "Your adherence to the cautions presented here will permit you to use MER/29 effectively in your practice."

Dr. Nestor of the FDA telephoned early the following morning. He said, as reported in a memorandum from Dr. Murray to Getman, that the FDA did not have sufficient pertinent facts to decide on the warning letter, that it feared side effects might be more severe and extensive than stated, and that the letter "might, therefore, be misleading and lead the physician to a false sense of security." The FDA wanted all facts available to the defendant "on the cases of eye changes, including the actual case records," a written copy of the proposed letter, and a statement that defendant had "supplied all toxicity data in animals and man, including that available to us from outside sources." Dr. Nestor had recounted discussions of MER/29 with several experts who were concerned with "numerous other serious side effects" which he enumerated, had expressed the FDA's feeling that they had enough evidence to suspend approval of the drug, and had told of advice from their statistician that Merrell's claims for significant lowering of cholesterol levels were not true. A meeting was arranged for October 26, and Getman directed that the FDA's requests be fully met. Meetings, some of them apparently rather stormy, were held on October 26 and again on November 2,29 17 and 21.

Meanwhile, on October 20 defendant wrote, as it was free to do, to clinicians who had been administering the drug at high dosages, telling them of the reported cataract cases, now four in number, and strongly urging discontinuance of the drug if hair or skin changes appeared. It discontinued placing new advertisements and instructed salesmen to return all samples and promotional literature other than the FDA approved labelling, cautioning them, however, not to mention that their samples had been recalled. Merrell also sought the advice of two independent clinical investigators and the professor who had chaired the Princeton conference as to whether MER/29 should be withdrawn from the market; all favored continued marketing with a strengthened warning.

The letter approved by the FDA on November 27 and sent on December 1 to all physicians and all the defendant's salesmen began by noting that while "comparatively few serious clinical injuries have been reported to date, their possible significance is emphasized by findings from animal studies." It next told of the four human cataract cases, one involving a patient receiving the recommended daily dosage of 250 mg., and added that "cataracts and corneal opacities have also been produced with MER/29 in animals." It said that slit lamp examinations were necessary for early detection of cataracts and thus were "indicated prior to and periodically during therapy." The letter went on with hair changes and skin changes, saying these might be related to each other and to cataracts, and recommended that MER/29 therapy be discontinued immediately on their occurrence. It also reported certain scattered cases of other adverse effects in humans and more in animals. While observing that "the side effects of all types reported to us to date total substantially less than one percent of the patients treated," including a number under therapy for more than a year and a few in excess of three, the letter "recommended that MER/29 be used only in patients who can be maintained under very close supervision and frequent observation" and never at more than 250 mg. per day. It concluded by saying that further studies were under way and requesting any information the physician could contribute.

Although plaintiff makes many other criticisms of defendant's conduct in late 1961 — alleged withholding of data as to other forms of toxicity in animals,30 an attempt to submit to the FDA an inadequate version of what Merck had reported, and deficiencies in its instructions to salesmen — we have discussed the most damaging evidence and the rest adds little to a resolution of the real issue, whether a jury could reasonably find that the defendant's proposed letter was a mere cover-up to permit continued sale of MER/29 without caring whether this would cause cataracts or not. Here, as in the case of defendant's decision not to alert the FDA or the medical profession as to the Merck report, everything depends on where one starts. If there was a fair basis for concluding that the combined effect of the rat and dog experiments and the reports from the Mayo Clinic had convinced defendant's management that MER/29 had a significant cataractogenic potential in humans, the question could certainly be answered in the affirmative. However, we find no evidence sufficient to warrant a conclusion that the management entertained any such belief. Here again it is all too easy to be misled by hindsight. The eye damage in rats and dogs under heavy dosage did not necessarily indicate that this would be encountered in humans, of whom many thousands more had by then taken the drug, and Mayo was not certain that their cases were attributable to MER/29. Neither is there evidence to suggest that defendant had lost confidence in the highly beneficial effects of the drug. Moreover, and this goes to a good deal of the case, some weight must be given to the human tendency to follow a course of conduct once decided upon even when considerations have appeared that would have led to a different decision at the outset, a tendency peculiarly strong when large investments of both effort and money have been made; the very fact of the initial decision importantly affects subsequent ones. See James, Psychology (The Briefer Course) § 17-22 (1961 paperback ed.). Although this tendency can be pushed to the point of recklessness, a court should be careful not to set the scale too low when a discovery of social utility is under review. A strong case of recklessness could have been mounted against Columbus had he returned to Palos with lives lost and nothing found.

Furthermore, while defendant's proposed letter would have served as a first warning, it was never suggested that it would be the last. Defendant's position all along was that it should be permitted to send out its warning letter immediately on the understanding that the letter would be subject to change as a result of further conferences and investigations. In fact, this might well have been better for the public than what the FDA accomplished. The essentials in which the December 1 letter differed from what defendant volunteered to send on October 19 were the recommendations that slit-lamp examinations be made before and during therapy and that MER/29 be used only under very close supervision and frequent observation, and the omission of the reassuring concluding paragraph. We cannot say that, on the basis of what it then knew, defendant's failure to propose so strong a warning could reasonably be found reckless or wanton.

Plaintiff's final suggestion is that MER/29 was taken off the market in April 1962 not because of the increasing reports of cataracts, culminating in that of the six year old boy at the Mayo Clinic, but because of a surprise inspection of the Merrell plant by the FDA which revealed the fabricated monkey data discussed some time back. Taking the evidence in the sense most favorable to plaintiff, the utmost shown is that the raid may have accelerated the decision by a few days; there is nothing to warrant an inference that but for this defendant would have continued marketing the drug in the face of increasing evidence of its cataractogenic qualities.

Plaintiff places heavy reliance on the argument that even though no one of the incidents we have analyzed would itself warrant a finding of recklessness, their combination permits an inference of a plan which not only heightens the significance of each item but would permit consideration of other acts we have not detailed including those of subordinates. We have no quarrel with this as a general principle; indeed we have said quite recently that, even in a criminal case, "The trier is entitled, in fact bound, to consider the evidence as a whole; and, in law as in life, the effect of this generally is much greater than of the sum of the parts." United States v. Bottone, 365 F.2d 389, 392 (2 Cir.), cert. denied, 385 U.S. 974, 87 S. Ct. 514, 17 L. Ed. 2d 437 (1966). Still the burden devolves on the court to see that when all the evidence is thus considered, there is enough to warrant the finding that the law requires. And here we think plaintiff fails. The thorough discovery by his industrious and able counsel has unearthed countless instances of carelessness and even of wilfulness by subordinate officials and of failure to exercise proper supervision and possible bad judgment by higher ones. Granted that few human endeavors would escape without blemish from such searching scrutiny, the picture is not a pretty one. But there was no proof from which a jury could properly conclude that defendant's officers manifested deliberate disregard for human welfare; what it shows as to this, apart from negligence in policing subordinates and a somewhat stiff-necked attitude toward the FDA, is rather that they were so convinced of the value of the drug both to the public welfare and to the company's finances that they maintained a sanguine view longer than prudence warranted.

Moreover, New York demands, as it might have to before punishing a defendant with fines similar to those imposed on a criminal charge, that the quality of conduct necessary to justify punitive damages must be "clearly established." Cleghorn v. New York Cent. & H. R. R. R., 56 N.Y. 44, 48 (1874). Cf. Hedrick v. Jebiley, 198 N.Y.S.2d 346 (N.Y. City 1960). As the Supreme Court has recently reminded us, the standard of proof "by clear, unequivocal and convincing evidence * * * is no stranger to the civil law." Woodby v. I. N. S., 385 U.S. 276, 285, 87 S. Ct. 483, 17 L. Ed. 2d 362 (1966), citing 9 Wigmore, Evidence § 2498 (3d ed. 1940). It would be hard to think of a situation more appropriate for invoking that standard than where the manufacturer of a new drug honestly believed to assist in prolonging human life is faced with claims for penalties by hundreds of plaintiffs running into millions of dollars, in addition to many millions more for damages sustained. We have little doubt that in such a case the New York Court of Appeals would require a judge to scrutinize the evidence in far closer detail before submitting punitive damages to a jury than it would on an issue of compensatory damages for negligence or breach of contract. If that prophecy should prove to be wrong, we would then be obliged to decide in future cases whether we nevertheless have power to impose a higher standard of proof for the award of punitive damages in a federal trial, cf. San Antonio v. Timko, 368 F.2d 983, 985 & n. 1 (2 Cir. 1966), as we surely would if we do.

The judgment as to compensatory damages is affirmed; the judgment as to punitive damages is reversed. No costs.

My brethren, disregarding the Ostopowitz case in which a New York court passed upon the same issues as those involved here, speculate on what the higher New York courts might hold.

At the very least, I would await the result of appeal in Ostopowitz.

Roginsky's fifty-five page petition for rehearing warrants a few additional ones from us.

While contending that we are concluded by the trial judge's decision in Ostopowitz v. Wm. S. Merrell Co., Supreme Court, Westchester County, New York, No. 5879/63, see fn. 3 to our opinion, now under appeal by the defendant to the Appellate Division, Second Department, Roginsky urges, as "the chief relief sought," that we "suspend" our decision until appellate proceedings in that case and another, Space v. Richardson-Merrell, Inc., now on trial in Broome County, New York, are concluded. We do not regard ourselves as bound by the rulings of a state nisi prius judge although we treat these with respect. Insofar as Fidelity Union Trust Co. v. Field, 311 U.S. 169, 61 S. Ct. 176, 85 L. Ed. 109 (1940), on which plaintiff relies, may have indicated such a duty of abnegation, King v. Order of United Commercial Travelers, 333 U.S. 153, 68 S. Ct. 488, 92 L. Ed. 608 (1948) and Bernhardt v. Polygraphic Co. of America, 350 U.S. 198, 204-205, 76 S. Ct. 273, 100 L. Ed. 199 (1956) id. at 209-212, 76 S. Ct. 273 (concurring opinion of Mr. Justice Frankfurter), have removed this, in favor of Professor Corbin's earlier view that when a federal court must determine state law, it should not slavishly follow lower or even upper court decisions but ought to consider all the data the highest court of the state would use. See Corbin, The Laws of the Several States, 50 Yale L.J. 762 (1941). Such is the established position of this court. See, e. g., Strubbe v. Sonnenschein, 299 F.2d 185, 188-189 (2 Cir. 1962); Merritt-Chapman & Scott Corp. v. Public Util. Dist. No. 2, 319 F.2d 94, 103 (2 Cir. 1963); and Evans v. S. J. Groves & Sons Co., 315 F.2d 335, 341-345 (2 Cir. 1963), a negligence case in which we predicted the result later reached by the New York Court of Appeals in Pfaffenbach v. White Plains Express Corp., 17 N.Y.2d 132, 269 N.Y.S.2d 115, 216 N.E.2d 324 (1966). In any event we in no way disagree with the Supreme Court Justice's ruling that "high moral culpability" may justify the award of punitive damages under New York law, a position based on the leading case of Walker v. Sheldon, 10 N.Y.2d 401, 404-405, 223 N.Y.S.2d 488, 491, 179 N.E.2d 497 (1961), which we also cited, p. 842. And we cannot say whether or not we disagree with his holding that Mrs. Ostopowitz met the burden required for submission of the issue to the jury since, as previously pointed out, slip opinion 1828 n. 16, we do not know what the record in that trial was and do not consider ourselves required to search it out. Indeed, in an answer to the petition for rehearing which we permitted to be filed, defendant asserts that the Ostopowitz record does differ in some important respects.

We are similarly unimpressed by plaintiff's alternative position. If Roginsky wished a truly authoritative ruling on state law, he should have sued in a state court. The suggestion that the federal courts should suspend decision after their time has been taken by a twenty day trial and an appeal requiring detailed examination of an appendix of more than 2000 printed pages and hundreds of exhibits comes with rather ill grace when advanced by counsel for the first time more than three months after the appeal was argued and only after an adverse decision, although the Ostopowitz ruling was reported a week after the argument. We think plaintiff would have been rightly outraged if, in the event of affirmance, defendant had now made a similar suggestion. Moreover, plaintiff has received what the jury considered full compensation for the cataracts, arrested at an early stage,1 caused by his taking MER/29, if indeed they were so caused, see p. 836 (and the other cases pending in the District Court and the New York courts will afford ample opportunity for punishment and deterrence if our legal views should prove ill-founded. Since, as we indicated in our opinion, at n. 16, the district court should not press the other MER/29 plaintiffs to trial if they prefer to await the final decision of similar cases by the New York courts, we fail to understand why our prediction of New York law will have the large impact on these actions that counsel fear. In any event this would not seem a matter in which Roginsky is concerned.

Plaintiff objects to our remarks on the problem of multiple punitive damage awards for serious negligence in the manufacture of goods or the rendition of services, claiming that this question has here arisen "when the MER/29 litigation is pretty well at an end."2 Our concern over multiple punitive damage awards was not, as counsel contend, solely about Richardson-Merrell or even drug manufacturers in general but about all manufacturers, suppliers and consumers of goods and services. Pages 839-841. Also, as plaintiff concedes, the only conclusion we drew from this analysis was "that the consequences of imposing punitive damages in a case like the present are so serious that the New York Court of Appeals would subject the proof to particularly careful scrutiny." A good discussion that has become available since our opinion was rendered states that "If one accepts the proposition that the consequences of punitive damages can be `momentous and serious'" — as can scarcely be denied where the defendant is exposed to liability to many plaintiffs — "then justice requires increasing the burden of persuasion of the plaintiff in a punitive damages action." Comment, Criminal Safeguards and the Punitive Damages Defendant, 34 U.Chi.L.Rev. 408, 417, 418 (1967).3

We have reviewed plaintiff's extensive criticisms of our summary of the evidence with the care required by the earnestness of the petition and by counsel's technique of referring to testimony and exhibits in a staccato manner that often creates an impression rather different than when the evidence is read. We find nothing to persuade us that our summary was in error.4 While counsel are quite right that we did not mention every item of evidence on which plaintiff had relied, as our opinion clearly stated at p. 844, we had carefully considered the additional items cited. The case was indeed a close one and we did not decide that plaintiff had "no proof" as his petition quite unjustifiably says but rather "no proof from which a jury could properly conclude that defendant's officers manifested deliberate disregard for human welfare," p. 850. After thorough reconsideration that remains our view;5 counsel tend to forget that even when the burden of persuasion is but the ordinary one of preponderance, the winner of a verdict is entitled only to inferences the evidence fairly supports.

Counsel's final point is that it may have been a mistake to try the case and argue the appeal on the basis that complicity by defendant's presidents or vice-presidents must be shown to warrant a punitive award, see p. 842, n. 17. They find support for the view that the acts of inferior supervisory employees may be enough in the brief opinion in Soucy v. Greyhound Corp., 27 A.D.2d 112, 276 N.Y.S.2d 173 (3d Dept. 1967), a case decided on the pleadings which we cited for a different point although it had not been cited to us and which we do not find very helpful on the instant question,6 and in General Motors Acceptance Corp. v. Froelich, 106 U.S.App.D.C. 357, 273 F.2d 92 (D.C.Cir. 1959), a case which although cited to us we did not discuss in view of the agreement of counsel, conceded in the petition for rehearing, "that officer participation must be shown," under New York law. Plaintiff asks us to direct the district judge to grant him a new trial on punitive damages at which he can present this theory or, alternatively, that we allow the judge to do so. We shall make no such direction. On the other hand, nothing in our judgment would prevent the judge from granting this extraordinary relief if in the exercise of a sound discretion he should think it to be warranted.

The petition for rehearing is denied.

I dissent.