LACOMBE, Circuit Judge
(after stating the facts). The trial judge charged the jury that the burden of proof rested upon the defendant (the insurance company) to show by a preponderance of evidence that “the fall preceded the fire”; that “this building did not fall as the result of fire.” Exceptions to the charge and to refusals to charge the converse of this proposition sufficiently present the question of correctness of this ruling. It will not be necessary to repeat the text either of the charge or of the requests. The trial judge construed the clause referring to a fall of the building as a proviso or condition subsequent defeating any claim of the insured. If it be such, no one here disputes the proposition that the burden of proving the happening of the subsequent condition would rest upon the insurer. The defendant, however, contends that the clause is an exception to the general liability assumed by the insurance company, and that, therefore, it was for the insured to show that the loss did not come within the terms of the exception. The general rule is well expressed by Earl, J., in Slocovich v. Insurance Co., 108 N. Y. 56, 14 N. E. 802:
“Where there is an insurance against a loss by fire, and it is proved or admitted that the property insured has been destroyed by fire, the loss is brought literally and exactly within the terms of the policy. If, in such a case, the insurance company claims to be exempt from paying the sum insured because there has been a breach of some condition contained in the policy, or the violation of some obligation or duty imposed upon the insured by the law or contract, the burden rests upon it to establish the facts which it thus relies upon as a defense to the claim under the policy.”
The diligence of counsel has presented a long array of authorities bearing upon this assignment of error. The question has been expressly decided in accordance with defendant’s contention in Pelican Ins. Co. v. Troy Co-op. Ass’n, 77 Tex. 225, 13 S. W. 980, and Insurance Co. v. Boren, 83 Tex. 97, 18 S. W. 484, and in accordance with plaintiff’s contention in Insurance Co. v. Bamberger (Ky.) 11 S. W. [814]*814595, in Blasingame v. Insurance Co., 75 Cal. 633, 17 Pac. 925, and in Insurance Co. v. Crunk, 91 Tenn. 376, 23 S. W. 140.
In the Texas case the policy contained the following provisions:
“(1) This company shall not be liable for any loss or damage by fire caused by means of hurricane. (2) If the building shall fall, except as the result of fire, all insurance of this company on it or its contents shall immediately cease- and determine.”
The fire occurred during or immediately following a severe hurricane, which at least partially blew the house down, and there was evidence tending strongly to show that thé fire had its origin in the breaking of a lamp by falling timbers. The court held:
“The provisions of the policy above noticed are exceptions to the general liability assumed by appellant, and the petition should have -averred that the fire did not occur from one of the excepted causes. This was necessary to show a cause of action, for the company did not insure against loss resulting from a fire caused by a hurricane, nor were its policies binding at all for a loss caused by fire occurring after the fall of the house, unless the fall was caused by fire.”
In the Kentucky case the policy contained this clause:
“This company shall not be liable under this policy for loss and damage if the building herein described, or any part thereof, fall, except as the result of fire.”
The jury were instructed that defendant was not liable for any loss or damage, if the building fell, unless the fall was the result of fire, and that:
“The burden is upon-the defendant to show by the evidence that the building, or such part thereof as fell, * * * did not fall as the result of the fire; and, unless the jury believe from the evidence that the said building, or such part thereof as fell, did not fall as the result of the fire, they should find for the plaintiffs.”
These instructions were approved by the appellate court.
In the California case the policy provided that the company should not be liable for “loss caused by the fall of any building insured, or containing property insured, by this policy, or by fire ensuing therefrom.” The complainant alleged that all the property was totally destroyed by fire, but it was not alleged that the loss ivas not caused by any of the excepted causes. The complaint was demurred to on the ground that it contained no such allegation, and the demurrer was overruled. On appeal the supreme court sustained the judgment overruling the demurrer, saying:
“One seeking to recover on an insurance policy must -aver the loss, and show that it occurred by reason of a peril insured against; but he need not aver the performance of conditions subsequent, nor negative prohibited acts, nor deny that the loss occurred from the excepted risks.”
In the Tennessee case the policy provided that:
“If the building, or any part thereof, fall, except as the result of fire, all insurance by this policy on such building or its contents shall immediately cease.”
The defendant demurred on the ground that there was no averment in the declaration that the building insured, or any part thereof, . [815]*815did not fall except as the result of fire. The demurrer was overruled, and the supreme court, affirming such decision, says:
“It is not necessary that it should hare averred the performance or nonperformance of conditions subsequent, nor to have negatived prohibited acts or excepted risks.”
The burden of proof has been held to rest upon the defendant under other similar clauses of fire policies; i. e. clauses restricting in some way the liability of the insurer. So, where it was provided that the policy should be void in the event of the insured effecting additional insurance (Clark v. Insurance Co., 9 Gray, 148; Russell v.'Insurance Co., 84 Iowa, 93, 50 A. W. 546); also where it is provided that the policy should be void if the property was allowed to remain vacant beyond a limited time (Bittinger v. Insurance Co., 24 Fed. 549); where there was a provision that the policy was executed by the agent aud delivered to the insured upon the condition that it should not become effective until it was approved by the home office (Young v. Insurance Co., 59 Conn. 41, 22 Atl. 32); and where it was provided that the company should not be liable to make good any loss or damage by fire which might happen or take place by means of any invasion, insurrection, etc. (Insurance Co. v. Reynolds, 32 Grat. 613). A clause to the effect that the insurer should not be answerable for loss by fire which should happen by any explosion is referred to in two cases cited by defendant (Hayward v. Insurance Co., 7 Bosw. 385, 2 Abb. Dec. 349, and St. John v. Insurance Co., I Duer, 373, 11 A. Y. 516) as “an exception to the general language of the previous clause, by which they promise to make good such loss or damage as shall be occasioned by fire.” But the point here raised was not before the court. It was conceded in both cases that the tire was the result of an explosion, and the word “exception” is used in the opinions, evidently, not in its technical sense, as contrasted with “conditions.” but as a convenient way of expressing the fact that the insured under such a policy would not be liable for all losses by fire. This seems clearly indicated by the sentence from II N. Y. 518:
“Hence a loss occasioned ¡by invasion, insurrection, riot, and the like has usually been found excepted in such policies; and although in this, and perhaps in policies generally, the exception in this respect is in terms of losses by fire, the clause would be equally definite and intelligible if those words were omitted in the clause stating the exception.”
The academic distinction between an exception and a proviso is thus stated in Bouvier’s Law Dictionary:
“An exception exempts absolutely from the operation of an engagement or an enactment; a proviso defeats (heir operation conditionally. An exception takes out of an engagement or enactment something which would otherwise be part of the subject-matter of it; a. proviso avoids them by way of defeasance or excuse.”
It may not be always easy to apply this distinction in practice (witness the conflicting decisions supra), but no especial help to the solution of tire problem is to be derived from cases holding, as do some of those cited on the brief, that, when the insurance is against “perils of the sea,” the burden is upon the insured to show [816]*816that such perils caused the loss; or that, where insurers engage to he responsible for partial loss only if it amounts to 5 per cent., or the ship be stranded, the insurer must show loss to the specified percentage, or that the ship was stranded; nor that, under bills of lading providing that the carrier should not be liable for loss by fire while in transit, plaintiff, if he seeks to recover for a fire resulting from the carrier’s negligence, must show that negligence affirmatively; nor that under a policy insuring against injuries effected through external, violent, and accidental means the burden is upon the plaintiff to show from all the evidence that the death wras caused by external, violent, and accidental means. Baker v. Insurance Co., 12 Gray, 603; Paddock v. Insurance Co., 104 Mass. 521; Insurance Co. v. Shaw, 2 Dill. 14, Fed. Cas. No. 14,366; The Neptune, 6 Blatchf. 193, Fed. Cas. No. 10,118; Whitworth v. Railway Co., 87 N. Y. 413; Kaiser v. Latimer, 9 App. Div. 37, 41 N. Y. Snpp. 94; Claflin v. Meyer, 75 N. Y. 260; Lamb v. Railroad Co., 46 ,N. Y. 271; Insurance Co. v. McConkey, 127 U. S. 661, 8 Sup. Ct. 1360.
A pertinent case cited on the brief of plaintiff' in error is Sohier v. Insurance Co., 11 Allen, 336. The policy in that case insured Sohier, in the language of the opinion—
“Against loss or damage by fire to the amount of $2,500 on Ms brick and slate building known as the ‘National Theater,’ situate on Portland street, Boston, Mass. This policy not to cover any loss or damage by fire which may originate in the theater proper.”
Some provisos against liability for loss by fire which happens by invasion, riot, and the like are in a later part of the policy. The clause in italics is written in the policy, the rest of the parts quoted being printed.
The opinion proceeds (the italics infra being our own):
“The first question raised by the bill of exceptions is whether the burden of proof was on the plaintiff to show a loss by fire which did not originate in the theater proper. This depends upon the construction given to the clause, ‘This policy not to cover any loss or damage by fire which may originate in the theater proper.’ If that clause can be regarded as a proviso, — that is, a stipulation added to the principal contract, to avoid the defendant’s promise by way of defeasance or excuse, — then it is for the defendants to plead it in defense, and support it by evidence, but if, on the other hand, an exception, so that the promise is only to perform what remains after the part excepted is taken away, then the plaintiff must negative the exception to establish a cause of action. It is not always easy to determine to which class — whether of provisos or exceptions — a particular stipulation belongs; and this one is certainly very near the line.. But, after careful consideration, the court are of the opinion that this was an exception to the subject of the contract, and that it put the burden of proof on the plaintiff. The qualification of the contract to which the parties agreed is not inserted with any technical formality or precision. But it is found between the statement of what is insured and the promise to pay in case of loss, in close connection with, and qualification of, the description of the subject-matter of the insurance. The provisos are set forth in a different part of the instrument. It thus seems to be a direct limitation of the risk against which insurance is effected. The difference would only be a formal one if, instead of the phraseology actually used, the language of the policy had been, ‘do insnre against loss or damage by fire not originating in the theater proper.’ It would illustrate the operation of the phrase in question, and show its effect as an exception, if we suppose it applied to the building insured. If the clause in the policy had been, ‘This policy not to cover any loss or damage by fire to the part of the building used as a theater proper,’ * * * this would [817]*817manifestly have been an exception from the subject-matter of the insurance. And it is in like maimer an exception to the risks taken by defendants, when, in, the same pari of the policy in which, they insure the risk of fire, and in the same connection, they state, in substance, that it is only fire which does not originate in the theater proper against which they insure.”
A similar question lias been before the courts in actions upon policies of life insurance, where the contract sued upon contained a clause that it was to be void if insured died of a disease induced or aggravated by intemperance, or was to be null anu void if he should die in, or in consequence of, a duel, or in violation of the laws of any nation, state, or province; or, where the clause read, “The self-destruction of the person [insured], whether voluntary or involuntary, and whether he be sane or insane at the time, is not a risk assumed by the company in the contract.” In each of these cases it was held that the clause contained a condition, and that it was for defendant to show its breach by a fair preponderance of proof. Van Valkenburg v. Insurance Co., 70 N. Y. 605; Murray v. Insurance Co., 85 N. Y. 236; Goldschmidt v. Insurance Co., 110 N. Y. 628, 17 N. E. 871. And a similar rule of construction has been applied in cases of accident insurance, where the contract provided that the insurance should not extend to cover accidental injuries or death caused by fighting or voluntary exposure to unnecessary danger, nor while the insured was under the influence of intoxicating drinks. See Jones v. Association (Iowa.) 61 N. W. 485, where the court says:
“Not one of these conditions was to happen prior to the time the contract between the assured and the company should become binding. * * * The situation then is this: That there was a valid contract of insurance when the policy issued, but it might thereafter, upon the happening of some of these conditions, cease to be enforceable. * * * They were each and all matters of defense available to the defendant, but, not constituting a part of the plaintiff’s case, the burden did not rest upon him either to plead or prove them in the first instance.”
And in Coburn v. Insurance Co., 145 Mass. 226, 13 N. E. 604, where there was a similar ruling, the court uses this language:
“Stipulations added to a principal contract, which arc intended to avoid the defendant’s promise by way of defeasance or excuse, must be pleaded in defense, and must be sustained by evidence. They are in the nature of provisos. Exceptions which leave the defendant liable to perform that which remains after the part excepted is taken away are to be negatived.”
Some remarks of the court in Freeman v. Insurance Co., 144 Mass. 572, 12 Y. E. 372, are most pertinent to the case at bar. The policy insured against bodily injuries “effected through external, violent, and accidental means, within the intent and meaning of this contract, and the conditions hereunto annexed.” After the principal clause followed five provisos and eight conditions. The second proviso was: “Provided, always, that this policy is issued and accepted subject to all the provisions herein contained or referred to.” The third proviso was: “That this insurance shall not extend to any bodily injury * * * when injury may have happened in consequence of voluntary exposure to unnecessary danger.” The first condition was: “The party insured is required to use all due diligence for personal safety,” etc. The last condition provided: “The provi[818]*818sions and conditions aforesaid, and a strict compliance therewith during the continuance of this policy, are conditions precedent to the making of this contract.” The question presented on appeal was whether the burden of proof was on the plaintiff to show that the insured used “all due diligence for personal safety and protection.” The court says:
“The rule of pleading in, declaring upon a contract which contains an exception, or a proviso, or a condition is stated in Com. v. Hart, 11 Cush. 130, 134, as follows: ‘If such an instrument contains in it, first, a general clause, and afterwards a separate and distinct clause, which has the effect of taking out of the general clause something that would otherwise be included in it, a party relying upon the general clause in pleading may set out that clause only, without noticing the separate and distinct clause which operates as an exception; hut, if the exception itself be incorporated in the general clause, then the party relying upon it must, in pleading, state it, together with the exception.’’ It is a general rule of the law of evidence that it is necessary for a party to prove the substance of facts which he is required affirmatively to aver in his pleading. I't is true that the policy in the case at bar only insures against bodily injuries effected by the means described ‘within the intent and meaning of this contract, and the conditions hereunto annexed,’ but this does not change the nature of the conditions. The condition we are considering is essentially an executory stipulation in the form of a condition that Murray shall use all due diligence for his personal safety and protection, and it is the breach of tills condition by Murray which the defendant sets up as a defense. We are not aware that it has ever been held that the introduction of the words we have quoted, or other similar words, into the principal clause of a policy of insurance, incorporates into this clause the conditions of the policy, within the meaning' of the rule of pleading we have stated; and in some of the decisions, where it has been held that the defendant must plead, or that the burden of proof is on him to show, that a representation was false, or that a stipulation contained in a condition had not been complied with, the policy contained these or similar words in the principal clause. In an action upon a policy which contains many provisos and conditions there is a practical wisdom which courts have recognized in compelling the insurance company to allege and prove the want of compliance with any particular proviso or condition on which it relies.”
Examined in the light of these authorities, the clause providing what shall happen in the event of a fall is not difficult of construction. It is not in that part of the policy which insures the risk, nor “in close connection with and qualification of the description of the subject-matter of the insurance,” but is placed with the other provisos, in a different part of the instrument. The mere location of the clause is, of course, not controlling, but it has been considered as of some weight in several of the cases cited supra. Nor is it to be construed as if it were removed from its position among the provisos, and incorporated with the clause descriptive of the subject-matter, by the mere use of the words, “except as hereinafter provided.” To give these words such an effect would be to incorporate with the descriptive clause all the provisos as to loss caused directly or indirectly by riot or invasion, or by neglect of the insured, or by explosion or lightning, or where there has been other insurance (not notified to company), or where manufacturing is carried on after 10 p.m.,or the building stands vacant, etc. The overwhelming weight of authority, as will he seen from the citations supra, is opposed to any such construction. The most important element, however, in determining whether a particular clause expresses a condition or an exception is the nature of -the clause [819]*819itself. What does this particular clause mean, “If a building, or any part thereof, fall, except as the result of Are, all insurance by Ibis policy on such building or its contents shall immediately cease”? .Manifestly, it does not merely provide that the insurer will not he liable for the particular variety of loss by fire which results from a fall. Lt stipulates for very much more, viz. that the contract, which it is expressly provided shall normally continue for a year, shall, in the event of a fall, absolutely cease and determine, so that, if a fall shall take place which in no way injures the property insured, and it he thereafter destroyed by fire happening otherwise; than by fall or from prohibited causes, the insurer is nevertheless not liable, because an event has happened which, by agreement of the parties, puts an end to the contract altogether. It is difficult to see how such a clause can be construed otherwise than as a condition subsequent.
To the further argument that the words in the descriptive clause, “while contained in brick building,” etc., made it necessary for the plaintiff to show that no fall had destroyed the integrity of the building, a sufficient answer is found in the brief of the defendant in error. A clause drawn expressly to cover the case of a building falling before a fíre has been inserted in the contract, and it is to he assumed that the whole intention of the parties on that subject is expressed in such clause.
Tiie circumstance that, the complainant alleged that the “fire did not happen by i;' * * reason of any of the canses excepted by the terms of the policy” did not change the situation in any way. Unnecessary allegations in a complaint need not be proved. Murray v. Insurance Co., 85 N. Y. 239.
It is next contended by defendant that the court erred in refusing to charge, as requested, that:
“The burden upon the whole case is upon the plaintiff to establish by a preponderance of evidence that the damage sustained was caused by fire, for this is the only risk it was insured against under the policy; and, if all the evidence, when considered together, leaves the jury in doubt as to whether the damage was caused by fire or not, their verdict must be for the defendant.”
Manifestly, this is an accurate statement of the law, and presumably the only reason why the trial judge did not include it in his charge was because the question it: refers to was completely overshadowed by the other and more important one which has just been discussed. There seems to have been no conflict of evidence at all on this branch of the case. There is no suggestion anywhere in the record that the “damage was not. caused by tire.” The opening page of the brief filed by plaintilf in error contains this statement: “On April 30, 1895, the property insured was damaged and destroyed by fire.” The only question in the case was whether the fall preceded the fire, thus terminating the contract, or whether it did not. There was no contention upon the trial that the agency which destroyed the property was not the fire. If the jury had been required to answer specific questions, and if they liad found, first, that the fall did not precede the fire nor result from it:, and had found, secondly, that the damage was not caused by fire, it would have been the duty of the trial judge, upon motion, to set aside such finding, since it would be not only against the weight [820]*820of evidence, but wholly unsupported by any testimony in the case. Under these circumstances it was clearly unnecessary to instruct the jury as to the academic question upon whom there was imposed the burden of proving that the damage to the property insured was caused by fire. The fact was conclusively proved, and no one questions the accuracy of such proof.
It is next contended that the court erred in refusing to charge the jury that:
“Even though the jury find that the fire in the building preceded the fall, they cannot award the plaintiff any amount for loss following the fall, unless they find that the fall was the result of fire.”
The amount of the loss was not disputed upon the proof, nor, except for the suggestion that the fall preceded the fire, and thus caused the catastrophe, was there any claim that any part of the damage was caused otherwise than by fire. The court had charged the jury that the decision of the case “turns upon your conclusion as to a single issue of fact”; that “the suit was brought to recover a loss which the plaintiff alleges it sustained by the burning of its stock of goods,” etc.; that “the plaintiff alleges that on the night of April 29, 1895, while this policy was in force, a fire occurred, in consequence of which its stock of merchandise was nearly wholly destroyed”; that, “according to the theory of the plaintiff, — and there seems to be no dispute about this upon the evidence, — the loss which the plaintiff is entitled to recover, if it is entitled to recover at all, is the sum of $7,744.77, with interest,” etc.;, that the “simple issue, therefore, for your consideration is that which is presented by this defense: Did the building fall as a result of fire, or did the fall precede the fire?” These parts of the charge were not excepted to, and, having thus substantially covered the point now presented as the proof left it, the trial judge was under no obligation to instruct again upon the same point in the precise language of the request.
The next group of assignments of error raises the question as to the propriety of allowing one of the witnesses, a civil engineer, and expert in heavy construction work, to read excerpts from scientific hooks when giving his testimony. The general proposition that scientific books are not to be read in evidence is a familiar one, and many citations from text writers and reported cases are found in the brief of the plaintiff in error. Nearly all the reported cases deal with medical works, and most excellent reasons for the application of the general rule in such cases may be found therein. But the rule is not of universal application. It would be a reproach to the administration of the law if it were so. Records of observations are undoubtedly secondary evidence, but, if all such records were excluded from the sources of knowledge available to a court of justice, it would frequently find itself unable to obtain information which was open to every individual in the community. It was been held repeatedly that standard life and annuity tables, showing at any age the probable duration of life, are competent evidence (Railroad Co. v. Putnam, 118 U. S. 554, 7 Sup. Ct. 1); and yet these tables show merely the deductions from records of past transactions, when [821]*821neither the record of the transactions nor the individual who has worked out the deductions is called to testify to the accuracy of his work, or to the conditions under which it was performed. So, too, almanacs, astronomical calculations, tables of logarithms, interest tables, weather reports, tables of the rise and fall of the tide, have been admitted in evidence. In an opinion approving of the admission of market reports, upon which the commercial world relies, is found the following pertinent suggestion of Judge Cooley:
“As a matter of fact, such reports, which are based upon a general survey of the whole market, and are constantly received and acted upen by dealers, are far more satisfactory and reliable than individual enfries, or individual sales or inquiries; and courts would justly he the subject of ridicule if they should deliberately shut their eyes to the sources of information which the rest of the world relies noon, and demand evidence of a less certain and satisfactory character.” Sisson v. Railroad Co., 14 Mich. 497.
The particular excerpts complained of in the case at bar are these: Certain reports of the United States department of agriculture, prepared under the direction of the chief of the division of forestry, contain tables which comprise the results of over 2,000'tests by the United States government of the crushing strength of different lands of timber, prepared expressly to increase the knowledge of timbers grown in this country for the benefit of merchants and dealers find, builders and engineers. The report is a recognized authority in the engineer’s profession. Prom the tables the witness read the “results of investigation on long leaf pine,” which was the kind of timber in the posts the cause of whose giving way was the subject of dispute. The next book produced was Kent’s Mechanical Engineer’s Pocketbook, — the last edition of 1896, — which, it is not disputed, is a recognized authority. “Every mechanical engineer,” says the wilness, “has it on his shelf.” Prom a table in this book, giving the crushing strength of timber, the witness read a statement of such strength; per square inch, of the kind of pine of which the posts were made. The third book is Johnson’s Strains in Frame Structures, also concededly a recognized authority. It contained similar tables, and a similar excerpt was read. That information of great value is obtained by multiplying such tests and tabulating the results is surely self-evident. Under the rule contended for, that valuable information would be available for the use of a court, of justice so long as the men who made the tests and prepared the tabulations were living and producible, but after their death or disappearance the information they had gathered would be lost to the court, although available for every one else in the community, and relied upon by engineers and builders whenever a new structure is in process of erection. Upon the precise point here presented the diligence of counsel has not succeeded in discovering a single authority. We feel, therefore, no hesitancy in so modifying* the general rule as to hold that, where the scientific work containing them is concededly recognized as a standard authority by the profession, statistics of mechanical experiments and tabulations of the results thereof may be read in evidence by an expert witness in support of his professional opinion, when such statistics and tabulations are generally [822]*822relied upon by experts in tlie particular field of the mechanic arts with which such statistics and tabulations are concerned.
It is further assigned as error that two witnesses (Bowman and Stanton) called by the plaintiff were allowed to state their opinions us to whether the fall preceded the fire or the fire preceded the fall. It is not objected that the witnesses were not experts, and precisely similar questions had been put by defendant to its own witnesses. It appeared, however, that neither Bowman nor Stanton saw the ruins until long after the fire, and defendant insists that their opinions were, therefore, incompetent. This does not necessarily follow. Although the ruins had been pretty well cleaned up, there were plenty of timber and columns lying down in the bottom, and it is safe to say that the indications afforded by the condition of the columns had much to do with the formation of an opinion. The lapse of time may have rendered the opinion of but little value, but on so much of the testimony as the record contains we are not prepared to say that the trial judge should have excluded it altogether.
The only remaining exceptions are to the refusal of the trial judge to allow defendant’s witnesses Cashman and Freel to express an opinion as to “how long a fire would burn in the building before the posjs would be weakened,” and as to “what time would elapse before fire and smoke would appear.” The hypothetical question intended to elicit this information contained, so far as the record shows, no indication as to whereabouts in the building the fire broke out. It is manifest that this is an important — probably the most important — element in the hypothesis, and, without it, any opinion, expert or other, would be mere wild guesswork. The trial judge correctly excluded it.
The judgment of the circuit court is affirmed.