MR. CHIEF JUSTICE CURETON delivered the opinion of the court.

The opinion of the Court of Civil Appeals, 260 S.W. 938, states the facts in this case quite fully, and we will not relate them further than may be necessary to make plain our conclusions.

Marvin Lee Francis held a life insurance policy issued by the International Travelers' Association, a mutual assessment accident insurance company, chartered under the laws of Texas, and upon his death the defendant in error, as guardian of the beneficiaries thereunder, brought this action. The policy was for $5000, payable in the event of death caused "solely and exclusively by external, violent and accidental means." The insured died on June 26, 1920, from Ludwig's Angina, due to an infection following the extraction of a wisdom tooth a few days before.

There is no clause in the policy which would prevent a recovery for an accidental death resulting by or from a surgical operation or from an infection. The by-laws of the association, however, at the time the policy was issued, and at all times subsequent thereto, contained exceptions and provisions which, if given effect, would prevent recovery. The face of the policy contained language which was intended to make the exceptions contained in the by-laws a part of the contract.

The first question for determination is whether or not the exceptions contained in the by-laws, but not specified in the policy, may be invoked to defeat recovery. We are of the opinion that the exceptions in the by-laws, in so far as they limit the contingency insured against, cannot be given effect, and are not to be regarded as a part of the contract. This is so because of the statute. Revised Statutes, 1925, Article 4797, which was in effect when the policy sued upon was issued, in part declares:

"Every policy or certificate issued by any such corporation shall specify the sum of money which it promises to pay upon the contingency insured against, and the number of days after the receipt of satisfactory proof of the happening of such contingency at which such payment shall be made. Upon the happening of such contingency, such corporation shall be liable for the payment of such amount in full at the time so specified, subject to such legal defenses as it may have against same." * * * (Italics ours.)

This statute plainly states that the policy or certificate shall specify the sum of money it promises to pay "upon the contingency insured against," and clearly requires a statement of the "number of days after the receipt of satisfactory proof of the happening of such contingency" at which payment of the sum named shall be made. These statutory requirements cannot be complied with except by stating the contingency insured against. The insistence here is that the specification of the contingency in the broad language quoted from the policy and the reference to the exceptions in the by-laws is a sufficient compliance with the statute. With this we cannot agree. Since it is evident that the contingency must be specified in the policy, then under the statute the whole of the contingency must be named, — that is, the contingency with all exceptions, if any. We therefore cannot look to the by-laws for a limitation on the general language used in the policy naming the contingency insured against. The statute governs and controls. The language of the policy directing the policyholder to look to the bylaws to ascertain the full or limited contingency under which the company agrees to pay, is in direct conflict with the statute. The legislative mandate is dominant and must prevail. Corporate bylaws cannot be made to repeal or supersede a statute. In fact, the statute which requires the specification of the contingency insured against must be regarded as entering into and forming a part of the contract, to the same effect as if embodied therein. The restrictive by-laws, in so far as they limit the right of recovery for accidental death named in the policy, are not effective, and pass out of the case. International Travelers' Association v. Branum, 109 Tex. 543, 212 S.W. 630; Eaton v. International Travelers' Association, 136 S.W. 817; Southland Life Ins. Co. v. Hopkins, 219 S.W. 254; Cooley's Briefs on Insurance (2d Ed.), Vol. 2, pp. 1102, 1138, 1139, 1141, 1143.

The next question is one of greater difficulty, but after careful consideration we have concluded that the Court of Civil Appeals was correct in holding that death in this case was by accidental means within the terms of the contract. The policy was payable upon death caused by "external, violent and accidental means." The trial court found that the death was accidental, but denied recovery because of the failure to give notice of death. The Court of Civil Appeals likewise found that death was accidental within the terms of the policy, and upon the conclusion that the notice of death was given as required, reversed and rendered the case in favor of the defendant in error. The occasion of the death was a dental operation, — the pulling of a wisdom tooth and its related treatment. The cause of death was an infection which produced Ludwig's Angina, a result so extraordinary and rare, and so unrelated to the surgical act performed, that it must be regarded as accidental. The drawing of the tooth and treatment following were of course purposeful and not accidental, but the infection was not the necessary or usual result of this purposeful act. It was extraordinary, unusual, and very rare. The proper surgical act, therefore, must have been accompanied by something unexpected, unforeseen, and improbable. This unforeseen, unexpected, and improbable thing was the injection of the pythogenic organism into the tissues. Just how they were injected the evidence does not show with mathematical precision; and in the nature of things can never be done in any case. These organisms do not make their presence known by ordinary methods of detection. That they were introduced into the tissues at the time of and as a part of the surgical act of extracting the tooth and treating the lacerated tissues, seems to be a logical deduction from the facts proven. The tooth was extracted by Dr. Tipton, a regularly licensed and practicing dentist of the city of Dallas. At the time of the extraction he examined the patient and found that there was no swelling of the gums, no redness, and no pus. The tooth was not drawn because diseased, but because rather deeply embedded in the gum. The fact is that there was no infection until the flesh was lacerated by the drawing of the tooth. The infection then followed so closely that no one has attempted to name a time for the beginning of the infection different from that when the tooth was drawn. Dr. Lott, the surgeon who operated in an effort to save the life of Mr. Francis, in part testified:

"Ludwig's Angina is due to or is a reaction following some kind of an infection. I do not know the name or names of any other germs that would produce this disease other than staphylococcus and streptococcus. I do not know that I exactly traced the source of the infection, but I got from him a history that indicated to my mind that he was suffering from the infection from or following the extraction of a tooth, and naturally I thought that the cause of the infection was the extraction of the tooth.

* * * * * *

"Now, as to whether Mr. Francis would have died when he did die, I think that he died from the infection resulting from the pulling of the tooth. * * * I naturally assume that the infection following the extraction of his tooth was the cause of his death, and toxemia or blood poison resulting in Ludwig's Angina."

Dr. Yancy, one of the attending physicians, in part testified:

"I went to see Mr. Francis at St. Paul's Sanitarium, and I got a history from him that he had been sick possibly for three or four days. * * * I made an examination of Mr. Francis' mouth, and when I had finished the examination I got from Mr. Francis a history of the fact that he had had a tooth extracted several days prior to that examination.

* * * * * *

"I am of the opinion that his condition was due to an infection caused from a tooth extraction, which was carried into the structure involved, which resulted in what medical science terms Ludwig's Angina.

* * * * * *

"It is my opinion that he would not have died unless the tooth had been pulled. It is my opinion now that the extraction of that tooth was the direct cause of that trouble.

* * * * * *

"This disease that I have testified about is known as Ludwig's Angina, and the nature of the disease is that it is an inflammation of the tissues beneath the jaw bone and about the floor and walls of the mouth, and it really follows or is the result of some infection. * * * I did not find any evidence of any source of infection other than where the tooth had been extracted that would have caused that condition."

The finding of the trial court as to the relationship between the extraction of the tooth and the infection is stated by plaintiff in error in its application as follows:

"In a trial before the court, the court found as a fact that on the 17th day of June, 1920, M. L. Francis had a wisdom tooth extracted by a dentist in the City of Dallas, and received as a result thereof an infection of streptococci which caused blood poisoning to set in which resulted in his death on June 26, 1920. The trial court further found that such an infection from the extraction of a tooth to the extent that it is fatal was a situation feared by such operation, but which rarely occurs."

The Court of Civil Appeals found:

"On June 17, 1920, the insured had a wisdom tooth extracted by a dentist in the City of Dallas. Infection from the operation resulted in Ludwig's Angina, from which he died. Infection was an unforeseen, unexpected, and unusual result in such cases, and Ludwig's Angina was unusual, unforeseen and unexpected in such cases, and death from this or these causes was unforeseen and an unusual result.

"The infection was caused either by the entrance of germs on the dental instruments used, or from cotton used by the dentist in his treatment, or from some other means; at all events, germs found their way into the tooth socket, did their deadly work, producing the unusual, unlooked-for, and unforeseen result, death."

There is no serious contention here but that the death of the insured was accidental; but the insistence is that the facts do not show that it was caused "solely and exclusively by external, violent and accidental means." In effect the insurance company here insists that there is a distinction between policies which insure against accidental death and those which insure against death by accidental means. This Court has heretofore recognized this difference in the meaning of insurance contracts. Bryant v. Continental Casualty Co., 107 Tex. 582, 589, 182 S.W. 673, L. R. A., 1196E. 945, Ann. Cases, 1918A, 517; International Travelers' Association v. Ross, 292 S.W. 193; Robinson v. Aetna Life Ins. Co., 276 S.W. 900. We have no disposition other than to follow what we regard as the established rule. That the infection resulting in the death of Mr. Francis was not the natural and probable consequence of the extraction of his tooth, was an effect which does not ordinarily follow and could not have been reasonably anticipated, was not the result of design, and was unexpected, is plainly shown from the uncontradicted evidence. This clearly makes the death an accidental one. Joyce on Insurance (2d Ed.), Vol. 5, Secs. 2863, 2863a. In order, however, for the recovery on the policy to be affirmed, we must conclude that the death was produced by accidental means. We have concluded that it was so produced. There can be no doubt but that the extraction of the tooth was the "means" of infection. This is so regardless of the question as to exactly how and when the pythogenic organisms entered the tissue, — whether by contact with the surgical instruments and dressings, and thence into the lacerated tissues, or whether from some other attendant source. The fact is there was no infection until the flesh was lacerated by the pulling of the tooth, and the infection followed so closely that we cannot in a practical consideration of the subject separate the time of infection from the time of laceration; nor have the doctors attempted to do so, as shown by their testimony above quoted. In fact, they regard the drawing of the tooth as the cause of the infection. If we consider the drawing of the tooth as the means of infection, then clearly that means was external, violent, and accidental. The authorities declare that accidental means are those which produce effects which are not their natural and probable consequences. Mr. Cooley in his work on Insurance declares:

"Whether or not the means is accidental is determined by the character of its effects. Accidental means are those which produce effects which are not their natural and probable consequences. The natural consequence of means used is the consequence which ordinarily follows from its use — the result which may be reasonably anticipated from its use, and which ought to be expected. The probable consequence of the use of a given means is the consequence which is more likely to follow from its use than it is to fail to follow. An effect which is the natural and probable consequence of an act or course of action is not an accident, nor is it produced by accidental means. It is either the result of actual design, or it falls under the maxim that every man must be held to intend the natural and probable consequence of his deeds. On the other hand, an effect which is not the natural or probable consequence of the means which produced it, an effect which does not ordinarily follow and cannot be reasonably anticipated from the use of such means, an effect which the actor did not intend to produce and which he cannot be charged with the design of producing, is produced by accidental means. It is produced by means which was neither designed nor calculated to cause it. Such an effect is not the result of design, cannot be reasonably anticipated, is unexpected, and is produced by an unusual combination of fortuitous circumstances; in other words, it is produced by accidental means." (Italics ours.) Cooley's Briefs on Insurance, (2d Ed.), Vol. 6, p. 5234.

Tested by the rule quoted, the real basis of which will appear in the discussion which follows, death in this case was by accidental means.

The things actually done by the dentist in this case were the normal and usual things done by dentists without fatal results. Moreover, each of these things was the expected thing to which all consented. In consequence, the expected result would have been that the gum would heal and the patient recover in the usual way in the usual time. That was the result which all had the right to expect, because that would have been the usual result. But in this instance another factor entered into the operation, an unusual, unexpected, and unforeseen factor, to-wit: the pythogenic organisms which produced Ludwig's Angina, and subsequent death. The mouth and gums were examined by the dentist, and showed no signs of infection prior to the operation. They were badly and perceptibly involved within three days' time, and in eight days from the extraction of the tooth the patient was dead. The evidence shows that the tooth extracted was a wisdom tooth, rather deeply embedded in the gum, and the dentist was compelled to go deeper into the tissues to extract it than is ordinarily the case. The evidence also discloses that on the Sunday following the extraction of the tooth on Thursday Mr. Francis was suffering from intense pain, but that the surface of the tissue did not show inflammation. This strongly indicates that the seat of the infection was in the deeper part of the gum, into which the forceps had been forced.

There is but one reasonable conclusion from the evidence, and that is that the organisms of infection entered the system by reason of the extraction of the tooth, and evidently at a time concurrent with the extraction or the subsequent treatment of the wound. Their entrance into the lacerated tissues at the time and under the circumstances was an unusual and unexpected thing, and must be regarded as a vis major, an Act of God, and as much of an accident as any other act within the usual meaning of the word. The vis major, the Act of God, in bringing about the infection, either as a part of or concurrent with the voluntary act of extracting the tooth, or the treatment of the wound thereafter, was the unknown factor which produced the unexpected and unforeseen result. Since the extraction of the tooth was the means which brought about the accidental infection, the extraction of the tooth itself becomes the accidental means of infection. The expressions we have quoted from Cooley above do not state that because of an injury we infer an accident, but the meaning is that where normal and ordinarily means are employed in the usual way, and instead of producing the usual, normal, and expected result, an altogether different and fatal thing happens, the result is an accidental one, because an unknown factor, unexpected and unforeseen, has entered into the circumstances, which the authorities call a vis major or an Act of God; and when the vis major is so connected with the means employed, as in the instant case, as to be a part and parcel of what would otherwise have been a voluntary act, then the means employed, though employed voluntarily, take color from the unknown and fatal factor and become accidental. United States Mutual Accident Association v. Barry, 131 U.S. 100, 33 L.Ed., 60; Bryant v. Continental Casualty Co., 107 Tex. 582; Horton v. Travelers' Ins. Co., 45 Cal.App. 462, 187 P. 1070; Interstate Business Men's Accident Association v. Lewis, 257 Fed. Rep., 241; Iowa State Traveling Men's Association v. Lewis, 257 Fed. Rep., 552; Lewis v. Ocean Accident and Guarantee Corporation, 224 N.Y. 18, 7 A. L. R., 1129; Nax v. Travelers' Ins. Co., 130 Fed., 985; Western Commercial Travelers' Association v. Smith, 85 Fed., 401, 40 L. R. A. 653; Townsend v. Commercial Travelers' Mutual Accident Association, 231 N.Y. 148, 17 A. L. R., 1001; Continental Casualty Co. v. Willis, 28 F.2d 707; Moore v. Fidelity and Casualty Co., 202 Cal. 465, 56 A. L. R., 860; Cameron v. Massachusetts Protective Association, 220 Mo. App., 780, 275 S.W. 988; Mutual Life Ins. Co. v. Dodge, 11 F.2d 487, 59 A. L. R., 1290; Bankers' Health and Accident Co. v. Shadden, 15 S.W.2d 704 (Texas. Civil App.); Home Benefit Association v. Smith, 16 S.W.2d 357 (Texas Civ. App.); Fort Worth Mutual Benefit Association v. Miller, 280 S.W. 338, (Texas Civ. App.).

The celebrated Barry Case, supra, from the Supreme Court of the United States, is a very fair illustration of the principle involved. Dr. Barry intended to jump without injury, but for an unknown and unseen cause he alighted heavily and not in a usual or expected manner. Some unknown factor entered into his act, which prevented the co-ordination of his muscles, with an unintended and unforeseen result. The Supreme Court of the United States held that the fact that when he jumped from the platform he alighted heavily was sufficient evidence of an injury brought about by external, violent and accidental means.

In the instant case the presence of the streptococci organisms on the instruments used in the operation, or in the subsequent treatment, or their presence in the air or surrounding medium, was as much an Act of God, and therefore accidental, as was the failure of the muscles of Dr. Barry to co-ordinate in the expected and usual way at the time of his injury; and as in the Barry Case, took from the act of the insured its voluntary character.

In the Bryant Case, cited above, the contingency insured against was sunstroke due to violent external and accidental means. The insured suffered sunstroke while voluntarily walking on the streets in the course of his business. This Court, in an opinion by Chief Justice Phillips, approved the principles which ruled the Barry Case, and held that the death of the insured was due to external, violent, and accidental means. The language of the opinion, 107 Texas, pp. 590, 591, applies with force to the question presented in the instant case, and plainly shows that the voluntary submission by the deceased to the dental operation was not a voluntary submission to an infection which produced death, since the presence of the organisms, either on the instruments used or in the surrounding air, was the vis major, uncontrolled by human action, which produced the fatal results.

The case of Horton v. Travelers' Ins. Co., 187 P. 1070, quoted from in the opinion of the Court of Civil Appeals, is exactly in point. There the insured voluntarily underwent a dental operation. Blood poison resulted from the unintentional use of infected instruments, from which death resulted. The Court held that death was due to external, violent, and accidental means.

In the three Lewis Cases, above cited, two from the Federal Court and one from New York, the actions arose from the same state of facts. The contingency insured against in the New York case was from "bodily injuries effected solely through accidental means." The death of Lewis, the insured, was due to inflammation of the brain, caused by an infection arising from voluntary puncture of a pimple on the lip. The Court of Appeals of New York held that death was due to accidental means. In part the court said:

"We think there is testimony from which a jury might find that the pimple had been punctured by some instrument, and that the result of the puncture was an infection of the tissues. If that is what happened, there was an accident. We have held that infection resulting from the use of a hypodermic needle is caused by 'accidental means.' Bailey v. Interstate Casualty Co., 8 App. Div. 127, 40 N.Y. Supp., 513, id., 158 N.Y. 723, 53 N.E. 1123; Marchi v. Aetna L. Ins. Co., 140 App. Div. 901, 125 N.Y. Supp., 1130, id., 295 N.Y. 606, 98 N.E. 1108. The same thing must be true of infection caused by the puncture of a pimple. Unexpected consequences have resulted from an act which seemed trivial and innocent in the doing. Of itself, the scratch or the puncture was harmless. Unexpectedly it drove destructive germs beneath the skin, and thereby became lethal. To the scientist who traces the origin of disease there may seem to be no accident in all this. 'Probably it is true to say that in the strictest sense, and dealing with the region of physical nature, there is no such thing as an accident.' Halsbury, Ld. Ch., in Brintons v. Turvey (1905) A. C. 230, 233, 2 Ann. Cas., 137. But our point of view in fixing the meaning of this contract must not be that of the scientist. It must be that of the average man. Brintons v. Turvey, supra; Ismay v. Williamson (1908) A. C. 437, 440, 77 L. J. P. C. N. S. 107, 99 L. T. N. S., 595, 24 Times L. R., 881, 52 Sol. Jo., 713. Such a man would say that the dire result, so tragically out of proportion to its trivial cause, was something unforeseen, unexpected, extraordinary, and unlooked-for mishap, and so an accident. This test — the one that is applied in the common speech of men — is also the test to be applied by courts." 7 A. L. R., 1130.

In the two cases in the Federal Court, the Circuit Court for the Eight Circuit made a similar holding, sustaining the right of recovery. In these cases the record showed the insured voluntarily punctured a pimple on his lip with a scarf pin, and the contingency insured against was death due to external, violent and accidental means.

In the case of Nax v. Travelers' Ins. Co., supra, death resulted from a self-inflicted knife cut, made by insured while trimming a corn, which was followed by blood poison. The court held that death was from an "accidental, external, and violent" injury.

In the case of Cameron v. Mass. Protective Association, cited above, the contingency insured against was death due "solely to external, violent, and involuntary causes" and by "accidental means." The insured cut or lanced a pimple on his arm, from which death resulted. The Kansas City Court of Appeals held death was due to accidental means," saying:

"We think there is no question but that if the insured, in lancing or cutting the pimple on his arm, unintentionally used an instrument having thereon an anaerobic germ, or germs, and that such germs were inserted into his system, that the occurrence was an accident within the meaning of the policy. He intended to cut himself, but the insertion of the germs into his system was unintentional." 275 S.W., pp. 990, 991.

In the case of Moore v. Fidelity Casualty Co., above cited, the Supreme Court of California, sitting In Banc, affirmed the principles of the Horton Case, which we have heretofore discussed, and held that illness and death of a professional nurse from contact with germs of communicable disease while in attendance upon a patient, constituted a bodily injury through accidental means.

In Townsend v. Commercial Travelers' Mutual Accident Association, supra, the New York Court of Appeals held that death from septic poisoning from the voluntary use of a hypodermic needle was due to "accidental means."

The other cases cited support the conclusion reached by us. We are aware that there are cases, like that of Ramsey v. Fidelity Casualty Co., 143 Tenn. 42, 13 A. L. R. 651, 233 S.W. 84, which would support a holding directly contrary to that made by us. In fact, there are two clearly defined lines of authorities on the question here involved. Caldwell v. Travelers' Ins. Co., 305 Mo., 619, 39 A. L. R., 61. The conclusion reached by us, however, is the one consistent with the previous decision of this Court in Bryant v. Continental Casualty Co., supra, and is, we believe, sound in principle, and in harmony with the best reasoned cases.

The opinion of the Court of Civil Appeals on all the questions involved is manifestly correct, and we deem further discussion unnecessary. The judgment of the Court of Civil Appeals is affirmed.