The plaintiff appeals from a judgment, dismissing her complaint after a trial to the judge, in an action, brought as beneficiary, to recover upon a contract of life insurance upon her son’s life. There are only two questions: first, whether the defendant insured the son at all; and second, if so, whether he was intentionally shot, in which event a provision for “double indemnity” did not apply. The judge made detailed findings, the substance of which, so far as they are material to this appeal, is as follows. One, Kelman, a solicitor for the defendant authorized to take applications from prospective customers and to give receipts for first premiums, after two preliminary interviews with Gaunt, the insured, on August 3d, procured from him the signed “application,” which is the subject of the action. This was a printed document of considerable length and much detail, the only passage in which here relevant we quote in full in the margin.1 The important words were: “if the Company is satisfied that on the date of the completion of Part B of this application I was insurable * * *600and if this application * * * is, prior to my death, approved by the Company at its Home Office, the insurance applied for shall be in force as of the date of completion of said Part B.” Number 12 of the answers which the insured was to make in the “application” was in the alternative; it read: “Insurance effective: (Check date desired) Date of Part B □ Dated of issue of Policy When Gaunt signed the application he had not checked either of these answers; but after he had delivered it to Kelman, Kelman checked the second, so that, as the “application” read, Gaunt was to be insured only from the issuance of the policy. The judge found that “Both Gaunt and Kelman intended that Gaunt should be covered from the date of the completion of the medical examination”; and that Kelman’s checking of the wrong answer’ “was due to. a mutual mistake on the part of Gaunt and Kelman.” ■

At the time of signing the “application” Gaunt paid the full first premium and Kelman gave him a receipt containing the words we have just quoted without substantial change: both the “application” and the receipt were upon forms prepared by the defendant for use by solicitors such, as Kelman. On the same day Kelman took Gaunt to the defendant's local examining physician who found him insurable under the rules and who recommended him for acceptance. Kelman delivered the “application” and the premium, and the physician delivered the favorable report, to one, Wholey, the defendant’s local agent for Waterbury, Connecticut, who prepared a report recommending acceptance, signed by himself and Kelman, which he sent with the “application” and thp physician’s report to. the “home office,” where the documents were received on the 9th. Since it appeared from the papers that Gaunt had been classified as “4F” in the draft because of defective eyesight, the “medical department” at the “home office” required another physical examination in Waterbury. This took place on the 17th; on the same day the local physician wrote to the “home office” again passing Gaunt; and on the 19th “a lay medical examiner” for the “medical department” at the “home office” approved the “application.” Nevertheless the “home office” on the 20th wrote to Wholey asking further information as to Gaunt’s classification in the draft; Wholey answered satisfactorily on the 24th by a letter received on the 25th; and on the 26th one of the “doctors of the medical department * * * approved” the application “from a medical standpoint.” The “home office” received news on that day of Gaunt’s death, and never finally approved the “application,” although the judge found that, if Gaunt had lived, it would have done so.

Gaunt left Waterbury on August 19th. He was going to the Pacific Coast or to Alaska in search of’work; he arrived at Chicago on the 21st; and.on the 24th he had reached Montevideo, Minnesota, where he was seen traveling in an “army bus”’ that had been loaded upon a flat car of a west-bound freight train. The only other occupant of this bus was one, Rasch, about whom nothing was learned except that he was later traced to the wheat fields of Wyoming as a casual worker. On the 25th. Gaunt’s body was found beside the westbound track of the railroad at Milbank,. South Dakota, with a hole in his head made by a 38 or 45 calibre bullet, which had! entered his right jaw near the ear and had come out at the top of his skull; and although the record contains no evidence on the subject, we may take judicial notice that this must have caused substantially instant death. There was blood inside and! outside the bus, and the bullet was found inside which had killed him. On the testimony the judge found that Gaunt had been-intentionally killed, which, as 'we have said, was an exception to the “double indemnity”’ provision covering “accidental death.” The plaintiff asks us to reverse this finding: i.e., to find it “clearly erroneous”; but we should not be warranted in doing so. Neither side contends that Gaunt killed himself ; the issue is whether Rasch killed him accidentally or intentionally; and upon that the plaintiff argues that the defendant had the burden of proof. We hold that the *601evidence justified the finding, even if it did have the burden. It is apparent that Rasch, after he had shot Gaunt, must have dragged him out of the bus and placed him beside the track; and that he then fled, obviously to escape detection. The most reasonable inference is that he did this, hoping that the train would move on and that his presence in the bus with Gaunt would not be remembered, for, although no one saw him in it, one witness had talked with him at Montevideo and had learned that he was travelling in. the bus. It is true that the blood stains inside the bus were in any event a tell-tale circumstance of which he must have been aware — although not the presence of the bullet — but they were nothing like as incriminating as the body itself would have been where it was shot. That Rasch should have pulled out his revolver, shot Gaunt while merely examining it, and then have gone so far to escape implication, while possible, seems to us most unlikely. The evidence might not satisfy a jury in a trial for homicide, but there was certainly enough to support the affirmative finding.

The first question is whether Gaunt was covered at all at the time of his death. Curiously, neither party has incorporated in the record “Part B,” and we do not know what was the date of its “completion.” If it was the approval “from a medical standpoint” as “advised by one of the doctors of the medical department,” it was not “completed” before Gaunt’s death. On the other hand the judge found that “Gaunt was, at the time of the completion of Part B, insurable in accordance with the rules of the defendant company for the plan and the amount applied for,” and that is consistent only with the understanding that “completion” was earlier than the 25th. The defendant has not argued to the contrary and we shall so assume. Thus the question becomes whether the words: “if the application, including Part B, is prior to my death, approved by the Company, at its Home Office,” must inescapably be read as a condition precedent upon the immediately following promise: “the insurance * * * shall be in force as of the date of the completion of Part B.” It is true that if the clause as a whole be read literally, the insured was not covered if he died after “completion of Part B,” but before “approval” ; and indeed he could not have been because there must always be an insurable interest when the insurance takes effect.2 Yet what meaning can be given to the words “as of the date of the completion of Part B” if that be true? The defendant suggests six possible “advantages” to the insured which will satisfy the phrase, “the insurance * * * will be in force,” (1) The policy would sooner become incontestable. (2) It would earlier reach maturity, with a corresponding acceleration of dividends and cash surrender. (3) It would cover the period after “approval” and before “issue.” (4) If the insured became uninsurable between “completion” and “approval” it would still cover the risk. (5) If the insured’s birthday was between “completion” and “approval,” the premium would be computed at a lower rate. (6) When the policy covers disability, the coverage dates from “completion.” An underwriter might so understand the phrase, when read in its context, but the application was not to be submitted to underwriters; it was to go to persons utterly unacquainted with the niceties of life insurance, who would read it colloquially. It is the understanding of such persons that counts ;3 and not one in a hundred would suppose that he would be covered, not “as of the date of completion of Part B,” as the defendant promised, but only as of the date of approval. Had that been what the defendant meant, certainly it was easy to say so; and had it in addition meant to make the policy retroactive for some purposes, certainly it was easy to say that too. To demand that persons wholly unfamiliar with insurance shall spell all this out in the very teeth of the language used, is unpardonable. It does indeed some violence to the words not to make actual “approval” always a condition, and to substitute a prospective approval, however inevitable, when the insured has died before approval. But it does greater violence to

*602make the insurance “in force” only from the date of “approval”; for the ordinary applicant who has paid his first premium and has successfully passed his physical examination, would not by the remotest chance understand the clause as leaving him uncovered until the insurer at its leisure approved the risk; he would assume that he was getting immediate coverage for his money. This is confirmed by the alternatives presented in the twelfth question; the insurance was to be “effective,” either when the policy issued, or at the “date of Part B”; there was not an inkling of any other date for the inception of the risk. It is true that in Connecticut as elsewhere the business of writing life insurance is not colored with a public interest;4 yet in that state, again as elsewhere, the canon contra proferentem is more rigorously applied in insurance than in other 'contracts,5 in recognition of the difference between the parties in their acquaintance with the subject matter. A man must indeed read what he signs, and he is charged, if he does not; but insurers who seek to impose upon words of common speech an esoteric significance intelligible only to their craft, must bear the burden of any resulting confusion. We can think of few situations where that canon is more appropriate than in such a case as this.

Situations very close aboard have arisen not infrequently, although the actual words have necessarily varied, so that it is hardly fair to say that any decision is quite on all fours. However, the important question is how far the condition of subsequent approval shall prevail over the promise of immediate coverage as soon as the insured has paid his premium and has passed his physical examination. A number of the decisions rely upon Insurance Company v. Young’s Administrator;6 but it really does not touch the issue. In the first place, Ho-mans, the insurer’s agent who took the premium, was apparently not authorized to bind the company at all, but only .to transmit the application to the home office; as a general agent, he was not one of those officers whom the insurer had authorized to “sign for” premiums. 23 Wall, at page 89, 23 L.Ed. 152. Be that as it may, Young did not die until after the company had rejected the application and had tried to substitute another which he never accepted. Northwestern Mutual Life Insurance Co. v. Neafus7 was a similar case. Those decisions are scarcely in point where the “application,” or the receipt does not say expressly, but only by implication, that the coverage shall begin when the applicant had been examined;8 but there are unquestionably decisions which excuse the insurer when it does.9 The law of New York is in doubt. The first case was Hart v. Travelers’ Ins. Co.,10 which was in favor of the insured and which the Court of Appeals affirmed, though without any opinion.11 Moreover, this was followed by Buono'v. Prudential Ins. Co.,12 which accepted it as authoritative. However, both these decisions were themselves soon followed by three others in the Appellate Divisions of the Fifst and Second Departments,13 all of which went the other way. It is true that they distinguished Hart v. Travelers’ Ins. Co., supra,14 by the language of the application, and Buono v. Prudential Ins. Co., supra,12 on the ground of estoppel; but with deference it is hard to see how all the decisions can stand, and despite the affirmance of Hart v. Travelers’ Ins. Co., supra, we should not be warranted in holding that the *603point is settled. On the other hand, in Duncan v. John Hancock Mutual Life Ins. Co.,15 Stonsz v. Equitable Life Assurance Society of United States,16 and. Western & Southern Life Ins. Co. v. Vale,17 the insured won; and indeed in Albers v. Security Mutual Life Ins. Co.,18 the court went further in his favor than we need go here. Thus upon a preponderance of the decided cases the answer is in doubt, and we cannot be sure how a Connecticut court would decide for the point has never come up in that state. Unaided as we are, we rest our decision upon the reasons which we have tried to set forth.

We are satisfied that the “double indemnity” clause did not apply. As we have already said, the finding that Gaunt was intentionally shot was not “clearly erroneous” ; and the question of law as to who had the burden of proof upon the issue does not arise. It does not appear that the judge placed the burden upon the plaintiff; and, if that was an error, the plaintiff has not proved that he committed it. She also argues that the defendant is “estopped” to set up the exception to the “double indemnity” provision where the killing was intentional. This she bases upon the fact that Kelman said nothing about it at the time, though he did say that suicide would avoid the policy altogether. This argument is too frivolous to deserve serious discussion.

Judgment reversed; judgment to be entered for plaintiff for $15,000.

(concurring).

I agree that the course of negotiations required and controlled by the insurance company was “unpardonable,” and am willing to concur in the decision for that reason. But I do not think we can properly or should rest upon the ambiguity of the company’s forms of application and receipt. Had this bargaining occurred between parties with equal knowledge of the business and on equal terms, there could be little difficulty in supporting the condition precedent that the “insurance,” i. e., the insurance contract or policy, could not “be in force,” i. e., take effect, until approved at the home office, and that then it dated back to an earlier time. Moreover, conditions of this general form are unfortunately still too customary for a court to evince too much surprise at them. There have been acute discussions of the legal problems involved; thus, most helpful is the article, Operation of Binding Receipts in Life Insurance, 44 Yale L.J. 1223.1 There receipts given for the payment of the first premium were held best divisible in two categories, one requiring approval as a condition precedent to the contract, in substance as here, and the other requiring that the company be satisfied that on the date of the medical examination the applicant was an insurable risk, and that the application was otherwise “acceptable” under the company’s regulations for the amount and plan of the policy applied for. The first form, it was said, was generally held to prevent the existence of a contract before acceptance, except with a few courts which found the provision too inequitable to support. The second, however, gave no difficulty where its reasonable requirements were afterwards found to have been met. A questionnaire to insurance officials showed an increasing trend towards the second or fairer form — a development warmly supported by the author. There was further the acute observation that use of the former form resulted in continuous litigation in a field of law where certainty was essentially indispensable, since it stimulated judicial intei-pretation to resolve the “ambiguity” against the company, followed by the latter’s renewed attempts to revise and refine the technical words.

Iffence a result placed not squarely upon inequity, but upon interpretation, seems sure to produce continuing uncertainty in the law of insurance contracts. Even though for my part I should feel con*604strained to concede the weight of judicial authority against our view,2 I think the considerations" stated are persuasive to uphold recovery substantially as would occur under the second form of contract stated above. I am somewhat troubled as to the state of local law in view of the stress in Swentusky v. Prudential Ins. Co. of America, 116 Conn. 526, 165 A. 686, upon the absence of unique features to insurance law. But that was. actually in another connection, a fact which I think justifies us in not here abdicating our judicial role for that envisioned by Judge Frank in Richardson v. Commissioner of Internal Revenue, 2 Cir., 126 F.2d 562, 567, 140 A.L.R. 705, of “ventriloquist’s dummy” as to state law.