dicial investigation in the ordinary way the question whether the fact concerning which inquiry was made, and an untrue answer given, was material to the risk. If it is in this manner found to be material, then the plain implication of the statute is that the usual penalty for breach of insurance condition and warranty shall follow, and the policy be avoided, whether the answer be made in good faith or not. If, however, the question untruly answered relates to something not found to be material to the risk, and if the answer is in good faith, then the breach of warranty works no prejudice to the insured or his representatives. If, though the question untruly answered relates to something not directly material to the risk, the untrue answer is made in bad faith,—that is, with a knowledge of its falsity, and for the purpose of misleading the company into the contract,-the implication of the statute is that the rule at common law shall prevail, and the policy shall be avoided. The statute has been construed by the supreme court of Pennsylvania, and our conclusions above stated are in accordance with the views of that court. Hermany v. Association, 151 Pa. St. 17, 24 Atl. 1064. In that case the court say (page 23, 151 Pa. St., and page 1064, 24 Atl.):

"This act has effected a change in life insurance contracts,-a much-needed change so far as some companies are concerned. The questions of materiality and good faith are ordinarily questions of fact, and therefore for the jury. They were certainly so in this case." "The evident purpose of this legislation was to strike down, in this class of cases, literal warranties, so far as they may be resorted to for the disreputable purpose of enforcing actually immaterial matters. It provides a rule of construction for the purpose of preventing injustice, and it is as much the duty of courts to enforce such rules as it is to administer the statute of frauds and perjuries."

The construction of a state statute by the highest court of the state is usually authoritative in courts of the United States. Burgess v. Seligman, 107 U. S. 20, 2 Sup. Ct. 10. And, even if it were otherwise, we should reach the same conclusion in this case. The court of appeals of Maryland has had occasion to construe this same statute, and has given it a like interpretation. Association v. Ficklin, 74 Md. 172, 21 Atl. 680, and 23 Atl. 197.

Having settled the construction of the statute, we come now to the questions of evidence. The circuit court was right in holding that within the scope of the question, "Have you your life. insured in this or any other company? (If so, give the name of each company and the kind and amount of the policy)," were not included Schardt's certificates of insurance in the Knights of Pythias and Royal Arcanum Mutual Aid Associations. It will be conceded that these associations, which are primarily for social and charitable purposes, and for securing efficient mutual aid among their members, are not usually described as insurance companies. That the certificate which they issue to a member, insuring upon certain conditions the payment of a sum certain to the member's representatives on his death, has mych resemblance in form, purpose, and effect to an insurance policy, is true; and, if we were called upon to give the application a wide and liberal construction

in favor of the insurance company, we might properly hold that the question embraced in its scope every association or individual contracting to pay money to one's representatives in the event of his death. Such a construction might be warranted by the probable purpose of the question to enable the company to judge how great a motive his life insurance would furnish the applicant for selfdestruction, or the fraudulent simulation of death. But we are here considering a contract and application drawn with great nicety by the insurance company, and framed with the sole purpose of eliciting from the insured full information of all the circumstances which the company's long experience has led it to believe to be valuable in calculating the risk. We cannot presume the company to have been ignorant of the fact that large numbers of persons have taken out life insurance in mutual benefit associations which are not ordinarily described as insurance companies, and that doubt has often arisen whether the contracts they issue are properly or technically described as life insurance at all. Insurance Co. v. Chamberlain, 132 U. S. 304, 10 Sup. Ct. 87. Having in view the wellestablished rule that insurance contracts are to be construed against those who frame them (Indemnity Co. v. Dorgan, 16 U. S. App. 290, 309, 7 C. C. A. 581, 58 Fed. 945; Insurance Co. v. Crandal, 120 U. S. 527, 533, 7 Sup. Ct. 685), and that any doubt or ambiguity in them is to be resolved in favor of the insured, we conclude that a certificate in a mutual benefit and social society was not within the description, "policy of life insurance in any other company." We are fortified in the conclusion by the fact that this contract is a Pennsylvania contract, and the courts of that state have uniformly held that mutual aid associations and insurance companies are so clearly to be distinguished that statutes applying to insurance companies and their policies do not have application to mutual aid associations, and the certificates of life insurance which they issue to their members. In Dickinson v. Ancient Order United Workmen, 159 Pa. St. 258, 28 Atl. 293, the defendant association sought to avoid its certificate on the ground of misrepresentation in the application. The plaintiff objected to the introduction of the application because it had not been attached to the policy in accordance with the Pennsylvania statute which forbade the introduction by an insurance company, in defense of a suit on its contract of insurance, of an application not attached to the policy when issued. It was held that the statute did not ap ply, because the defendant association was not an insurance company, but belonged to the distinctly recognized class of organizations, known as "benevolent associations." See, also, Association v. Jones, 154 Pa. St. 99, 26 Atl. 253; Com. v. Equitable Ben. Ass'n, 137 Pa. St. 412, 18 Atl. 1112; Com. v. National Mut. Aid Ass'n, 94 Pa. St. 481; Lithgow v. Supreme Tent (Pa. Sup.) 30 Atl. 830; Theobald v. Supreme Lodge, 59 Mo. App. 87; Sparks v. Knight Templars, 1 Mo. App. Rep'r, 334. It is true that in other states it has been held that such associations are within the description of "insurance companies," and that the contracts they make are properly termed

"policies," as those terms are used in the statutes of such states. State v. Nichols, 78 Iowa, 747, 41 N. W. 4; Insurance Order v. Lewis, 12 Lea, 136; Assurance Fund v. Allen, 106 Ind. 594, 7 N. E. 317; Com. v. Wetherbee, 105 Mass. 159; Sherman v. Com., 82 Ky. 102. In this conflict of authority, we must lean towards the decisions of the state courts of that state, according to the laws of which we must construe this contract, and, for the reasons already given, hold that certificates of membership in mutual benefit benevolent associations were not embraced in the question asked by the company in that state.

We now come to the questions of evidence with respect to the $5,000 policy in the New York Life Insurance Company which Schardt omitted in his answer to the question concerning other insurances. It is first insisted for the plaintiff below that his answer was not untrue. He was asked if he had other policies in other companies, and, if so, to state the companies and amount. It is urged that when he gave three such policies the question was answered correctly, and that his failure to give the fourth policy did not involve a false statement, but only left the answer incomplete, but true in everything stated. Several cases are cited to the point that such an answer is not a misrepresentation. In Perrine v. Society, 2 El. & El. 317, the applicant was asked what was his profession, and he answered that he was an "esquire." In fact, he was an ironmonger. It was held that there was no misrepresentation here, but, at the most, only a concealment or falsehood by implication; that the answer was true, as far as it went. The same ruling was made by the court of appeals of New York in Dilleber v. Insurance Co., 69 N. Y. 256. There the applicant was asked to state the physicians he had consulted in the last 10 years. He answered that he had consulted Dr. Paine 9 years before. In fact, he had also consulted another physician. It was held that, the answer being true as far as it went, there was no breach of the warranty; that the answer was full and true. We do not think that these cases can be supported. In Insurance Co. v. Raddin, 120 U. S. 183, 7 Sup. Ct. 500, the supreme court held that, where the answers to questions were obviously incomplete, the insurance company, by failing to inquire further before issuing the policy, waived any right to complain of such incompleteness; but the court clearly indicated its view that if such an answer was apparently complete, but in fact was otherwise, it was a false answer, and a breach of the warranty of its full truth. Towne v. Insurance Co., 7 Allen, 52, 53; London Assurance v. Mansel, 11 Ch. Div. 363; Bliss, Ins. (2d Ed.) 189, 190; Phil. Ins. §§ 550, 565, 567. The answer to such a question contains the necessary implication that there is no other insurance than that stated, and, if there is other insurance, it is as false as if the existence of other insurance were expressly denied. As already stated, any answer to a question, though concerning a matter not material to the risk, if made with intent to deceive the insurance company, would avoid the policy. Hence, even assuming that the question of other insurance was found by the jury to

be not material to the risk, the company still had a complete defense, if it could show that the answer had been made in bad faith. The intent of Schardt in omitting the New York Life policy, therefore, became a substantial issue, and evidence relevant to show his intent should have been admitted. The company offered to prove that, in answers to similar questions in applications for other policies, he had made answers equally untrue. We think this evidence was relevant and competent. It might have been forcibly argued on behalf of the defendant that Schardt had a motive to suppress the amount of other insurance, in the fear that, if the defendant knew all his then insurance, it would prompt inquiry into his purpose in carrying insurance in an amount out of proportion to his regular income of $1,500, upon which he was obliged to support a family, and would lead to a rejection of his application. And if the defendant could show a similar suppres sion of the same fact in the two applications for the later policies for $25,000 each, made within three months thereafter, when the same motive may be supposed to have been present, it would properly strengthen the argument that his suppression of the extent of his insurance in this case was with intent to conceal and deceive. Such evidence would have a tendency to show that his omission in the three cases was not by accident, but by design. It is a wellestablished rule of evidence that, where the issue is the fraud or innocence of one in doing an act having the effect to mislead another, it is relevant to show other similar acts of the same person having the same effect to mislead, at or about the same time, or connected with the same general subject-matter. The legal relevancy of such evidence is based on logical principles. It cer tainly diminishes the possibility that an innocent mistake was made in an untrue and misleading statement, to show similar but different misleading statements of the same person about the same matter, because it is less probable that one would make innocent mistakes of a false and misleading character in repeated instances than in one instance. Thus, where one was on trial for selling skimmed milk for fresh milk, in violation of the statute, it was held competent to show other instances of similar sales on other days by the accused about the same time, because, if he sold skimmed milk in repeated instances, it was rendered more probable that he knew its character in each instance. He might have made the mistake once, but not frequently. Bainbridge v. State, 30 Ohio St. 264. So, in this court, where the question was of the defendant's motive and knowledge in making statements concerning the character of a silver mine, we held it competent to show an elaborate and fraudulent scheme to mislead, not the plaintiff, but another, into the purchase of the mine, although the scheme was concocted and carried into attempted execution at least two years before the statements and sale to the plaintiffs. It was the circumstances that the acts related to the sale of the same mine, and that the motive for its sale might be presumed to continue, that removed the objection based on remoteness in point of time.

Mining Co. v. Watrous, 9 C. C. A. 415, 61 Fed. 163. Judge Lurton, in delivering the opinion of the court in that case, said:

"It is not, in such a case, essential that these former acts of fraud were not contemporaneous with the transaction under inquiry. If they were frauds of like character, and especially if they concerned former efforts to sell the same property, they are admissible. Remoteness in point of time may weaken their evidential value, but will not ordinarily justify exclusion."

Judge Lurton cites in support of this view Ross v. Miner, 67 Mich. 410, 35 N. W. 60; Hoxie v. Insurance Co., 32 Conn. 21; Rafferty v. State, 91 Tenn. 655, 16 S. W. 728; Bottomley v. U. S., 1 Story, 136, Fed. Cas. No. 1,688; Jordan v. Osgood, 109 Mass. 461; Castle v. Bullard, 23 How. 174; Butler v. Watkins, 13 Wall. 457; Insurance Co. v. Armstrong, 117 U. S. 598, 6 Sup. Ct. 877; Blake v. Assurance Soc., 4 C. P. Div. 94. It would seem clear from the foregoing that the objection made by counsel for the plaintiff that the other false statements of other insurance were too remote in point of time is not tenable. But it is suggested that the fact that the instances sought to be proven were subsequent to the instance in issue destroys their relevancy, because the fraudulent intent present in them might have been formed after an innocent mistake. This possibility, of course, affects the probative force of these subsequent instances to show fraud, but we do not think it makes them inadmissible. In Wood v. U. S., 16 Pet. 342, the question was whether there had been fraud in invoicing importations under the customs revenue law. It was objected that, while similar undervaluations in other importations prior to the one in issue might be admissible, still it was error to admit such undervaluations in later importations. To this, Mr. Justice Story, speaking for the court, said:

"The other objection has as little foundation, for fraud in the first importation may be as fairly deducible from other subsequent fraudulent importations by the same party as fraud would be in the last importation from prior fraudulent importations. In each case the quo animo is in question, and the presumption may equally arise and equally prevail."

For the error in excluding evidence of false statements concerning other insurance in the subsequent policies, the judgment herein must be reversed. The case will doubtless be tried again, however; and it becomes our duty, therefore, to examine and decide other questions made upon this record by the defendant which must, of necessity, arise again on the second trial.

At the trial the defendant introduced witnesses who had been long engaged in the life insurance business, and was permitted by the court to ask them whether the facts concerning which it was either admitted or claimed that Schardt had made untrue statements, and the fact of his embezzlements which he did not disclose, were material to the risk; but the court declined to permit an answer to the question whether, by the usage and practice of all insurance companies, such facts were regarded as material. This latter ruling of the court was excepted to by the defendant company. The question of evidence thus presented has been before the courts of England and America in many different phases, and the decisions present a bewildering conflict of authority. In the