SANBORN, Circuit Judge.
At tbe close of the plaintiff’s evidence tbe circuit court directed tbe jury to return a verdict for tbe defendant. This writ of error is sued out to reverse the judgment upon this verdict. There was but one question of fact in tbe case, and that was whether or not a policy of insurance, which, confessedly, prior to that date, did not cover the boiler which exploded, was so modified February 24, 1892, by a verbal agreement, that it did cover it on March 21,1892, when it exploded.
The Laclede Fire-Brick Manufacturing Company, the plaintiff in error, was a corporation engaged in manufacturing fire brick at Cheltenham, in St. Louis. In May, 1891, it had seven steam boilers, and no more, on its premises. It made a written application to the Hartford Steam-Boiler Inspection & Insurance Company, the defendant in error, "a corporation engaged in the business of insurance, to insure these seven boilers against explosion; and on May 8, 1891, the defendant issued to the plaintiff-its policy upon the seven specified boilers, insuring each of them, in the sum of $30,000, for three years, against explosion, provided the pressure of steam did not exceed 100 pounds per square inch on six of the boilers nor '85 pounds per square inch on the seventh, whén the explosion should occur. The defendant was a Connecticut corporation. The policy provided that it- should not bind the defendant unless it was countersigned by C. C. Gardner, its general agent, and it was so countersigned. The plaintiff alleged in its petition that on February 24, 1892, this policy was modified by the agreement of the defendant so that, without the payment of any additional premium, it thereafter covered nine boilers of the plaintiff, whenever only seven were exposed to the pressure of steam, and that on that day the defendant caused the two additional boilers to be inspected, and reported them sound. The defendant, by its answer, denied the agreement *353of modification, and alleged that the plaintiff applied to it for such a modification, and it caused the boilers to be inspected, and found that the attachments were not completed, and the boilers were not sound, and declined to insure them until the defects were remedied, and the attachments made.
In the conduct of its business, the defendant caused the boilers it insured to be inspected and tested, before taking risks upon them, and every few months during the continuance of the risks. The men it employed to make these tests were called “inspectors.” On February 24, 1892, Mr. Eickhoff, one of the defendant’s inspectors, tested the boiler which exploded, at the request of the plaintiff, and told the plaintiff’s engineer that it was sound; but it was not then inclosed, nor were the attachments to it made, and it was not subjected to steam pressure until March 19, 1892. There was no testimony that the general agent or any officer of the defendant made, or was informed of, any modification of the policy; but the agreement concerning it was claimed to have been made with the inspector Eickhoff, and is based on the following testimony: Mr. Green, the president of the plaintiff, testified: That Mr. Eickhoff was one of the defendant’s inspectors. That he first met him at some time while the plaint iff was insured by the defendant in the sum of $10,000 by a former policy. That, at some time subsequent to this first meeting, Mr. Eickhoff told him that he could not afford to inspect the plaintiff’s boilers every few -months for the money he was getting for it, and suggested that the plaintiff raise its insurance to $30,000, and he replied: “All right. Go ahead, and make it thirty thousand dollars.” That at that time he told Mr. Eickhoff that he might put in additional boilers, and that, if so, they would be auxiliaries or duplicates; that the risk would be no greater; that six boilers would he running for two weeks, and seven for the next two weeks, and the rest would he idle, — and Mr. Eickhoff said, “All right,” the insurance by the policy would certainly cover this risk, as it was less every two weeks, and no greater at any one time. That he supposes they talked this a dozen times, lie testifies that, after ihis talk, Mr. Eickhoff either brought to him, or left on his desk, the written application for the policy on the seven specified boilers then on his premises; that this application was not filled out by any one connected with the plaintiff; that he signed the application for the plaintiff, as president, and delivered it to Mr. Eickhoff, who took it away, and brought back, and delivered to him, the policy of May 8,1891; that he never had any conversation with any agent or officer of the defendant, except Mr. Eickhoff, regarding his insurance, until after the explosion. He testifies that he was the president of the Helmbacher Steam-Forge Company, which operated a rolling mill, and was not insured by the defendant; that shortly after the policy was issued he bought the two additional boilers, one of which subsequently exploded, for that company; that he did not then intend to have them insured, but intended to use them at the rolling mill; that before buying them he asked Mr. Eickhoff, as his friend, to inspect them for him, so that he might *354know whether or not they were a good purchase, and he did so, and reported them sound; that about six months later he asked him to inspect them for him again, and he did so, and reported that they needed some little repairs, which he caused to be made; that he then bought them for the plaintiff, and took them to Cheltenham, and put them up, and that when the brick walls were up about four feet he sent for an inspector; that Mr. Eickhoff responded to the call, and he asked him to inspect them again, and have the city inspector test them, and “he says to me: ‘You telephone to your engineer to let the water out of the boiler. Our inspection is' good enough. I will go out this afternoon and tell him so,’ which he did. During this conversation, I asked him if he considered those boilers insured, and he says, ‘I do.’ I says, ‘Will you attend to this business for me.’ * * * He says:. ‘I will. It is all right. Go ahead.’ He told me the boilers were insured, and they were in good shape; to ‘go ahead and put them walls up. The boilers are all right.’” He testifies that this was all the talk there was about insuring these boilers, at the time they were put' up; that this was at the continuance of a six-months talk; that he and Eickhoff had had a dozen conversations' about it; and that Eickhoff may have said the insurance would cover these boilers three or four times before that. George R. Blackford, the secretary of the plaintiff, testified that Mr. Eickhoff solicited Mr. Green to take the policy, and that, when the question of insuring the two additional boilers came up, he “said that the boilers would be insured under that policy;” that the second day after the explosion he went with Mr. Green to see Mr. Gardner, the general agent of the company; that Mr. Gardner said that the boilers' were not insured under that policy, and “Mr. Green then told him that it was his understanding with Mr. Eickhoff that the boilers were insured under that policy;” that at this interview Mr. Gardner stated that the boilers were inspected February 24,1892; that they were not then complete, and imperfections were pointed out, so that they were not in condition to be passed; that he showed them a copy of an inspection report of that date in the book of inspectors’ reports, and said that a copy of that report had been sent to the plaintiff, and that he would send it another copy; that the next day the plaintiff received a copy of such a report, which bears date March 24, 1892, has the words, “Take notice,” “Repairs' ordered,” on one side of it, and states that the boilers are defective, dangerous, and incomplete; and that the plaintiff never received any copy of such a report before that day, and had never asked for any inspection after the attachments of the boilers were completed.
Judges of the federal courts are not required to submit a case to a jury merely because there is some evidence in support of the case of the party who has the burden of proof, but it is their duty to instruct the jury to return a verdict against the party in any case in which they would be compelled to set aside a verdict in his favor, if rendered. Commissioners v. Clark, 94 U. S. 278, 284, and cases cited; Gowen v. Harley, 6 C. C. A. 190, 56 Fed. 973, 980, and cases cited. Was this evidence of such a character that it would warrant a jury *355in finding a verdict in favor of the plaintiff? This question must be determined by a consideration of the conversation of February 24, 1892, in the light of the surrounding circumstances. Unless that talk constituted a contract of modification of the policy, it was never modified. There is nothing in the evidence regarding the inspection of the boilers and the report of Eickhoff on that day, or the action of the defendant upon his report, that even approaches proof of a contract. It appears that in the course of its business the defendant-caused boilers to be inspected before it insured them, to see whether or not it was wise to do so, as well as after they were insured, to see whether or not it was safe hi continue its insurance; so that the fact that they were inspected tends as strongly to show that they ■were not insured, but an application to insure them had been made, as it does to show that their insurance had been effected. Nor does the fact tha t, on this inspection, Eickhoff reported to the plaintiff or to the defendant that the boilers were sound or unsound tend to establish any contract or modification of the policy, because he would have made one report or the other, whether such a contract had been consummated or not. His report to one or both parties that they were sound and safe to insure could not modify the policy of insurance, or make a contract of insurance, and certainly his report that they were unsound could not. Nor could the act of (lie defendant in sending, or failing to send, a copy of this report to the plaintiff, make such a contract, or prove that it had or had not been made. Hence, we are relegated to a consideration of the conversation of February 24,1892, to determine whether or not there was any contract of modification or of insurance made subsequent to the date of the policy. Before considering the effect of this conversation, it is well to note the character of the contract the plaintiff seeks to base upon it. Mr. Green testifies that, in the talk before he signed his application for the policy, he told Mr. Eickhoff, and the latter admitted, that, if nine boilers were so used that only.seven were exposed to the steam pressure for two weeks, and then the two idle ones, and four of those in use before were exposed to the same pressure for two weeks, and so on alternately, the risk of explosion would be no greater from the nine thus used than it would be from seven boilers constantly in use. This was a demonstrable mistake. The risk of the explosion of steam boilers is varied far more by the number of square inches exposed to the pressure than by the length of time the pressure continues. There is by no means twice as much risk that a boiler will explode in two hours, days, or weeks, under a pressure of a hundred pounds of steam, as there is that it will explode in one hour, day, or week. If it can withstand the pressure for any appreciable time, the risk that it will explode, under proper care, for many years, is comparatively light. But the risk that one of two boilers of equal size and strength will explode under a given pressure in a given time is twice as great as that either one will explode, because there are twice as many square inches of surface exposed to the pressure, and each inch must be able to resist it. This risk of explosion is analogous to the risk of the breakage of an iron chain sustaining a heavy weight. That risk increases in propor*356tion to the number of links in the chain, because every link must sustain the weight, while it is but slightly affected by length of time, short of that which would produce crystallization or stratifica-cation of the iron, because, if the chain will stand the strain one moment, it will do so indefinitely. If these boilers were able to stand the 100 pounds pressure to the square inch specified in the policy for one week, the risk during the three years named in the policy would be but slightly, if at all, diminished by permitting them to be idle two-ninths of the time, while the risk of explosion from nine boilers of equal strength would be two-sevenths greater than that from seven. The fact that one of these additional boilers did explode the first day it was exposed to full steam pressure shows that the risk of explosion from the exposure of additional surface to thq pressure was far greater than that of continuing the pressure on the surface that had been exposed for months. It follows that the contract of modification, if made, insured the plaintiff, in the sum of $30,000, for two years and three months, against the explosion of two old boilers, before their gauges were attached or their connections made, and several weeks before they were put in use. So radical a change in this contract ought not to be inferred, so burdensome a liability ought not to be imposed on this defendant, unless there is substantial evidence that at least the minds of Eickhoff and G-reen met upon and agreed to it.
Turning now to the conversation of February 24, 1892, between these two men, and bearing in mind their previous conversations, the condition of their minds, and the circumstances surrounding them, let us see what the evidence is that they agreed that the policy should be thus modified. They started, in their negotiation for the original policy, and still continued, in the erroneous belief that it made no difference in the risk whether seven boilers, in use all the time during the term of the policy, or nine boilers, each of which was in use seven-ninths of the time, were insured under the policy; hence, they both supposed that the modification pleaded by the plaintiff was immaterial to the defendant. Not only this, but before the policy was issued, at least a dozen times, Eickhoff had declared to Green that the policy on the seven boilers would cover all the additional boilers he might acquire after the date of the policy, if but seven were used at a time. After the policy was issued, Green says, Eickhoff continued to talk with him about it; that he talked about it a dozen times; that he may have said three or four times before the conversation of February 24th that the insurance would cover the two additional boilers; and that the talk of February 24th was at the continuance of these conversations. Is it not plain, from this testimony, that both these men were of the opinion that the policy, as it was, covered the two additional boilers, without any modification ? It seems so to us, and the subsequent conversations and acts of these men confirm this view. February 24, 1892, Green sent, not for Eickhoff, particularly, but for some inspector, not to insure his boilers, but to inspect them so that they would pass the city inspection. Eickhoff happened to be the inspector who responded to the call. Green immediately requested him, not to insure these boilers, but to inspect them, and *357have the city inspector test them. A long conversation followed about the method of inspection, about emptying the boilers, which were then full of water, and the necessity of calling the city inspector, which resulted in Eickhoff’s agreeing to make the inspection for himself and the city also. The only reference to the insurance in this entire conversation, according to Green, was that he asked Eickhoff if he considered the boilers insured, and he said he did. Green then asked him if he would attend to this business for him, and he said he would; that it was all right; that the boilers were insured, and were in good shape. The “business” that Eickhoff agreed to attend to — the “business” that Green sent for some inspector to attend to- — was the inspection and testing of these boilers, and not their insurance. The statement of Eickhoff that he considered them insured, that they were insured, was but a repetition of the opinion he had always expressed, — that the policy on the sevén boilers insured all the boilers the plaintiff could acquire, if only seven were used at a time. That this is the true construction of this talk is evidenced by the testimony of the secretary, Blackford, who said that what Eickhoff said was “that the boilers would be insured under that policy,” and that when, after the explosion, the general agent, Gardner, informed him and Mr. Green that these additional boilers were not insured by the policy, Green replied, not that he and Eick-hoff had agreed that the policy should be changed so as to insure them, but “that it was his understanding with Mr. Eickhoff that the boilers were insured under that policy.” In this entire record, the only suggestion of any contract modifying the policy, or insuring the additional boilers, subsequent to the date of the policy, appears in the pleadings, and is attributable to the superior knowledge and ingenuity of counsel. The plaintiff alleges such a contract. The answer denies it, but admits that an application for such a change was made. The evidence discloses neither contract of modification nor application for a contract of modification. Green and Eickhoff entered upon their negotiations under, two mistakes, — a mistake of fact, suggested by Green, that the risk of explosion from nine boilers in use seven-ninths of the time for three years was not greater than the risk of explosion of seven boilers constantly in use for that term; the other, a mistake of law, based on the opinion of Eickhoff that a policy insuring seven specified boilers would cover all the additional boilers the assured might acquire subsequent to its date, if he exposed but seven to steam pressure at a time. Laboring under these mistakes, Eickhoff gave it as his opinion, after the policy issued as before, that it covered the additional boilers the plaintiff acquired, and that he considered them insured by that policy. Green accepted that construction. In their view, it would have been a futile act to modify or change the policy, because they thought it was itself sufficient to accomplish their purpose.
It is true that a written contract may be modified by a subsequent oral agreement, and that a contract of insurance may be made by parol. But it is nevertheless true that the contract here in question was one of considerable magnitude, — one involving f30,000; that the customary method of modifying policies of insurance and of making *358contracts of insurance for long terms is by written agreements, and that, in most cases where oral contracts of insurance are made, they are initial, temporary contracts, to continue only until they can be embodied in a policy; and that the method pursued by the defendant when this policy was issued was to issue a written policy upon a written application.' The fact' that this talk was 26 days before the explosion, that no steps had been taken by either party meanwhile to put any contract of modification or of insurance in writing, and no demand had been made by the plaintiff for any such evidence of its contract, strongly indicates that no such contract was ever made. In Head v. Insurance Co., 2 Cranch, 127, 168, Chief Justice Marshall, in delivering the opinion of the supreme court, said:
“A contract varying a policy is as much an instrument as the policy itself, and therefore can only be executed in the manner prescribed by law. The force of the policy might, indeed, have been terminated by actually canceling it; but a contract to cancel it is as solemn an act as a contract to make it, and, to become the act of the company, must be executed according to the forms in which, by law, they are enabled to act.”
In Farley v. Hill, 11 Sup. Ct. 186, the supreme court declared the positive testimony of two witnesses to the existence of an oral agreement that but one witness denied, insufficient evidence to prove it, in view of the inherent improbability that such a contract would exist without some written evidence of it.
In. our opinion, the plaintiff’s evidence in this case was not only insufficient to establish the important agreement of modification of the policy or of insurance of the additional boilers he alleges, but, in the absence of the admission of the application in the answer, it fails to show that, after the issuance of the original policy; either Gfreen or Eickhoff ever contemplated, or suggested to each other, the making of any such agreement. This conclusion disposes of this case, for it is conceded, and is' too well settled to warrant more than the statement of the proposition, that the opinion Eickhoff expressed, or, if it could be so called, the “promise” he made, before the policy issued, that it would cover all after-acquired boilers, while but seven were in use, was merged in the written contract evidenced by the policy, and was not available to the plaintiff in this action either as a representation, an agreement, or an estoppel. Insurance Co. v. Mowry, 96 U. S. 544, 547, 549; Hudson Canal Co. v. Pennsylvania Coal Co., 8 Wall. 276, 290; Insurance Co. v. Lyman, 15 Wall. 664; Thompson v. Insurance Co., 104 U. S. 252; Pearson v. Carson, 69 Mo. 550; Tracy v. Iron-Works Co., 104 Mo. 193, 16 S. W. 203; Insurance Co. v. Neiberger, 74 Mo. 167; Lewis v. Insurance Co., 89 Conn. 100.
There is another reason why the judgment below should be affirmed, and that is that there is no sufficient evidence in this record that the inspector had authority from the defendant to modify the policy, or to make a supplemental contract of insurance in its behalf, to warrant a verdict against it. To maintain the negative of this- proposition, the authorities in support of the following propositions are cited: Where an insurance company defends against a loss on the ground that the policy is forfeited by a false representa*359tion in the application of the insured, it is in some cases an answer to this defense, in the absence of fraud, that the insured told the agent who solicited the insurance the truth, and the agent wrote the falsehood into the application himself. In such cases the false statement becomes the statement of the company, and not of the insured. Insurance Co. v. Wilkinson, 13 Wall. 222, 225; Insurance Co. v. Baker, 94 U. S. 610; Insurance Co. v. Mahone, 21 Wall. 152; Insurance Co. v. Chamberlain, 132 U. S. 304, 312, 10 Sup. Ct. 87; Insurance Co. v. Snowden, 58 Fed. 342; Kausal v. Association, 31 Minn. 17-21, 16 N. W. 430; Deitz v. Insurance Co., 31 W. Va. 851, 8 S. E. 616. An agent who is authorized to agree on terms of insurance may make a preliminary oral contract that the insurance to be evidenced by a policy shall commence from the date of the verbal contract. Insurance Co. v. Colt, 20 Wall. 560, 568; Insurance Co. v. Shaw, 94 U. S. 574; Fames v. Insurance Co., Id. 621; City of Davenport v. Peoria M. & F. Ins. Co., 17 Iowa, 276; Audubon v. Insurance Co., 27 N. Y. 216; Fish v. Cottenet, 44 N. Y. 538; Angell v. Insurance Co., 59 N. Y. 171. A corporation that holds one out as its agent in a particular business is bound by his acts within the scope of his apparent authority, although his real authority may be more limited, unless notice of the limitations is brought home to the party affected by his acts. Insurance Co. v. McCain, 96 U. S. 84; Griggs v. Selden, 58 Vt. 561, 5 Atl. 504; Walsh v. Insurance Co., 73 N. Y. 5. These propositions and authorities, however, do not rule this case. There is no attempt here to avoid the policy, and no question of false representation or description in it, or in the application on which it is based. Both plaintiff and defendant admit and count upon the existence and validity of the original policy. It goes without saying that the fact that an agent assumed to do an act is no evidence of his authority to do it, where that authority is questioned. Lohnes v. Insurance Co., 121 Mass. 439, 441; Bush v. Insurance Co., 63 N. Y. 531. This case is barren of evidence that Mr. Eickhoff ever had any real authority from the defendant to make or modify contracts of insurance on its behalf. It does appear that he was not the general agent of the company at St. Louis empowered to 611 out and countersign policies, and that one Gardner was. It does appear that his tide was not “insurance agent,” but “inspector,” and that his general business was inspecting and testing boilers. There is no evidence that he ever made, or agreed on the terms of, or solicited, any contract of insurance for the defendant, except the policy in question; so that his apparent authority to make and modify contracts of insurance for the defendant rests entirely upon his acts at the time of the issuance of that policy. What authority did he appear to have, from that transaction? He solicited the plaintiff to take ⅞30,000 insurance from the defendant. He told the plaintiff that this insurance would cover the seven boilers it then had, and all the boilers it ever acqtiired, if it used, but seven at a time. If this statement was, as we believe, his opinion of the legal effect of a policy on seven boilers, it certainly could not establish his authority *360to construe the contracts' of the defendant. If it was, as plaintiff claims, a promise, it was repudiated by the issuance of> the policy. When he solicited the policy, Eickhoff did not make, or assume to make, any oral contract of insurance for the defendant. He did not agree, or assume to agree, that the term of insurance should commence from the time of that conversation. So far as the evidence discloses, he went away without any discussion of, or agreement concerning, the time when the insurance should commence, the length of the term, the amount of the premium, or any of the other terms that were finally embodied in the policy, with the single exception of the amount. He subsequently brought to the plaintiff a written application for $80,000 insurance on the seven boilers only. There is no evidence that he filled out that application. The evidence is that the plaintiff did not, and there it stops. The plaintiff signed it. Eickhoff earned it away, and brought back the policy signed by the officers and the general agent of the defendant at St. Louis, but Eickhoff’s name does not appear on it. That policy did not insure the plaintiff against the explosion of boilers acquired subsequent to its date, if but seven were in operation; and if Eickhoffs statement was a promise, and not an opinion, the policy was in itself a repudiation of the talk of Eickhoff, and a notice to the plaintiff that the defendant did not recognize his authority to make terms of insurance. Hot only this, but the fact that no policy issued on the oral application of the plaintiff to Eickhoff, but a written application signed by the plaintiff, containing different terms from those named in the oral negotiations, was required, before the company would act at all, was complete notice and conclusive pi’oof that the defendant not only did not hold him out as having authority to make contracts for it, but did not even recognize his authority to receive oral applications ’for insurance, on which it would act. This was all the evidence that this inspector had any apparent authority even to make or modify contracts of insurance on behalf of the defendant, and it was insufficient. Insurance Co. v. Mowry, 96 U. S. 544; Morse v. Insurance Co., 21 Minn. 407; Healey v. Insurance Co., 5 Nev. 268, 274; Lohnes v. Insurance Co., 121 Mass. 439, 441. In Insurance Co. v. Mowry, supra, Mr. Justice Field, in delivering the opinion of the supreme court, said of a general agent of the insurance company, who had authority to take applications, countersign policies, and collect premiums:
“There is nothing in the record which shows that the agent was invested with authority to make an insurance for the company. In representing himself as an agent, he only solicited an application by the assured to the company for a policy.' That instrument was to be drawn and issued by the company, and it shows on its face that the authority to the agent was limited to countersigning it before delivery, and to receiving the premiums.”
In Morse v. Insurance Co., supra, an action was brought on an oral contract made by a soliciting agent at the time he obtained á written application, to the effect that the insurance should commence immediately, and a policy should be subsequently issued. The evidence was that the year before the agent had obtained a like application from the plaintiff, and made a like agreement, and *361the company had issued a policy for a term commencing a,t the date when the application was obtained, but there was no evidence that the company was notified of the agreement made by the agent. The plaintiff had a verdict, and the supreme court of Minnesota set it aside on the ground that the evidence of the agent’s authority was insufficient to warrant it.
How much stronger the evidence of the apparent authority of these agents to make insurance contracts was, than is the evidence in this case of the authority of this inspector, appears from what has already been said. Moreover, there is no evidence in this case that Eickhoif was the clerk or general representative of the general agent, in conducting his insurance business. The utmost stretch of his apparent authority reaches no further than to make him the occasional agent of the general agent to solicit an application, deliver it to him, and carry back the policy. The functions of such an agent cease with the delivery of the policy. From an apparent authority so limited, no authority to subtract from, add to, or modify the written contract of the insurance company can he inferred. Healey v. Insurance Co., 5 Nev. 268, 273; Putnam Tool Co. v. Fitchburg Mut. Fire Ins. Co., 145 Mass. 265, 269, 13 N. E. 902; Kyte v. Assurance Co., 144 Mass. 43, 46, 10 N. E. 518; Lohnes v. Insurance Co., supra; Tate v. Insurance Co., 13 Gray, 79, 82; Wilson v. Insurance Co., 14 N. Y. 418; Hoffman v. Insurance Co., 32 N. Y. 409; Walton v. Insurance Co., 116 N. Y. 317, 322, 324, 22 N. E. 443; Mitchell v. Insurance Co., 51 Pa. St. 402, 411. The judgment below is affirmed, with costs.
CALDWELL, Circuit Judge
(dissenting). There was abundant evidence to go to the jury, and from which they might rightfully have found, that Eickhoif was an agent of the defendant, and clothed with authority to modify or extend the original contract of insurance, and that such a contract was in fact made. His agency is not denied in the answer, but, by necessary implication, admitted. In its answer the defendant says:
“And defendant states the truth and the fact to be that at some time in the month of February, 1892, the plaintiff did apply to the defendant for insurance upon two additional boilers to the seven boilers mentioned in said policy of insurance, and did request the defendant to add said two boilers to said seven boilers, and did request defendant to agree that said policy of insurance issued to It in the month of Mareh, 1891, should cover said two additional, as well as said original seven, boilers. ⅜ * ⅞”
The testimony shows that all the dealings in reference to this insurance business — those relating to the policy that was issued in 1891, as well as those relating to the supplementary or ancillary *362agreement for the insurance of the two additional boilers — were had between Mr. Green and Mr. Eickhoff. Mr. Green was asked:
“Q. Who did you have your dealings with about this matter of this application? A. Mr. Eickhoff. Q. Did you see anybody but Mr. Eickhoff about it? A. No, sir. Q. Did you ever have any [negotiations] with Mr. Gardner at all? A. None at all. Q. Did you ever see any other representative of this company, either in taking out this policy or putting in this auxiliary battery of boilers, except Eickhoff? A. No, sir. Q. How did you come to ask Eickhoff to come and inspect these boilers for you? A. Well, he was doing my insurance.”
There is not a word of testimony to the contrary of this in the record, but much more to the same effect. Reading the paragraph of the answer I have, quoted in the light of this evidence, it is perfectly clear that the application which the answer admits was made “to the defendant” to insure the two boilers was the application made to its agent Eickhoff, for Mr. Green testifies positively that no such application was ever made to any one else. But what ought to be conclusive on the question of Eicklioff’s agency and authority to make the contract, as well as the fact that he did make it, is the circumstance that, when the plaintiff’s evidence tended .strongly to establish both these facts, the defendant declined to put either Eickhoff or Gardner on the stand to deny them. Whether Mr. Eickhoff was an agent of the company to effect insurance, and whether he did insure these two boilers, were facts peculiarly within the knowledge of himself and Mr. Gardner. It is a well-settled rule of evidence that when the proof tends to fix a liability upon a party who has it in his power to offer evidence of all the facts as they existed, and rebut the inferences which the proof tends to establish, and he declines to offer such proof, the jury are at liberty to presume that the proof, if produced, instead of rebutting, would support, the inferences against him. Railway Co. v. Ellis, 4 C. C. A. 454, 54 Fed. 481. In Starkie on Evidence (volume 1, p. 54), it is said:
“The conduct of the party in omitting to produce that evidence- in elucidation of the subject-matter in dispute which is within his power, 'and which rests peculiarly within his own knowledge, frequently affords occasion- for presumptions against him, since it raises strong suspicion that such evidence, if adduced, would operate to his prejudice.”
This rule is applicable to criminal cases (Com. v. Webster, 5 Cush. 295, 316; People v. McWhorter, 4 Barb. 438), and it is difficult to perceive why it should not be applied to an insurance company.
The bill of exceptions shows that the precise ground upon which the case was taken from the jury by the circuit court was "that there was no evidence to show that Mr. Eickhoff, the defendant’s inspector, had any authority from the defendant to change or modify the policy of insurance sued on.” The case comes here upon an exception to that particular ruling, and that is the only question discussed by counsel for plaintiff in error. It is now well settled that, when a given state of facts is such that reasonable men may fairly differ upon the question, the determination of the matter is for the jury. It is only where it appears that all reasonable men would agree that the evidence was insufficient to warrant a verdict that the court is justified in withdrawing a case from the jury. Rail*363way Co. v. Ives, 144 U. S. 408, 417, 12 Sup. Ct. 679; Railway Co. v. Ellis, 4 C. C. A. 454, 54 Fed. 481. Upon this question of fact, one judge of this court holds that the evidence was sufficient to warrant the jury to find a verdict for the plaintiff, one that it was not, and one is in too much doubt to express an opinion. . When this is the attitude of the appellate court, the cause ought, plainly, to be sent back to be tried by that tribunal appointed by the constitution to try the facts.
The remaining question is, did the agent Eickhoff make the ancillary contract to insure the two boilers? It is well settled that, unless prohibited by. statute or other positive regulation, a valid contract of insurance can be made by parol (Insurance Co. v. Shaw, 94 U. S. 574), and that an ancillary agreement, such as was made in this case, is binding, without any written memorial of it, and that such an agreement is not within the statute of frauds (Insurance Co. v. Colt, 20 Wall. 560). Whether the majority of the court mean to be understood as asserting that no such contract was made, or that, though made, it was void by reason of a mutual mistake of the parties, is not very clear, from the opinion. The supposed mistake, as stated by a majority of the court, consisted in Mr. Green and Mr. Eickhoff agreeing “that, if nine boilers were so used that only seven were exposed to the steam pressure for two weeks, and then the two idle ones and four of those in use before were exposed to the same pressure for two weeks, and so on alternately, the risk of explosion would be no greater from the nine thus used than it would be from seven boilers constantly in use.” As I understand the proposition of the majority, it is that the risk of explosion of two boilers of equal size and strength, when one is used one week and the oilier the next week, and so on, alternating, through the year, is just as great as if both boilers were used at the same time continuously through the year. This proposition is original, in this case, with the majority of the court. It is not found in the pleadings, was not raised in the court below, and is not alluded to in the briefs of counsel. It is not sound. It ignores the wear and tear and gradual deterioration and weakening that result from constant use. It takes no account of a change in the atmospheric conditions, and of the fact that boilers frequently explode from causes which are unknown, and cannot be ascertained. It is undoubtedly true that the view of the majority of the court on this question never suggested itself to the minds of Mr. Green and Mr. Eickhoff, although the latter is a skilled boiler inspector. But, if the proposition be true, it was a mistake of fact, or rather a "want of that scientific knowledge possessed by the majority of the court, relating to a matter about which both parties had equal opportunities of knowing what the fact was; and, there being, confessedly, no fraud on the part of either, the validity of the contract is not affected by the mistake. The consideration for the ancillary contract was ample. Mr. Green set up the boilers, and- incurred a large expense which he would not have done but for the agreement and assurance of the agent Eickhoff that they were insured, and that he might rest easy on that subject. The contract for the an-*364ciliary insurance was completed after the written policy was issued. The written policy bears date March 8, 1891, and the supplementary agreement was made in February, 1892. The answer admits that “some time in the month of February, 1892, the plaintiff did apply to the defendant for insurance upon two additional boilers;” and there was abundant evidence to go to the jury that such insurance was effected. When Mr. Green called on Mr. Gardner' after the loss, the latter did not then pretend that Eickhoff did not have authority to make the contract, or that the contract was not in fact made, but his claim was that “he didn’t think he was liable for anything that happened there, because it wasn’t bgcked on the back pf the policy;” evidently laboring under the impression that the company was not bound by the contract, because it had not been reduced to -writing. The testimony in this case shows that Mr. Green is a very prudent and careful business man, and that he had no thought of putting up these boilers without first having them inspected and insured. He did everything that a prudent man could do to effect that object. He would not purchase the boilers until he knew the defendant would insure them, and for that reason he applied to Mr. Eickhoff, the defendant’s agent and inspector, to inspect them before he purchased thein. And it was only upon being assured by the defendant’s agent that the boilers were sound, and that if he purchased them the defendant would insure them, that he concluded the bargain for them. When they were set up, and before casing them with brick, he had them inspected by the defendant’s agent and inspector, with a view to insuring them; and after the inspection he was assured by the defendant’s agent, Mr. Eickhoff, over and over again, that they were insured, and he laid out money upon the faith of that assurance. When a loss occurred, payment was refused, not because Eickhoff was not an agent,- or that no contract had been made, but because the contract “wasn’t backed on the back of the policy;” and, when the company is sued, it sets up various other defenses, which are obviously afterthoughts. The contract .of insurance should be characterized by the utmost honesty and good faith on the part of the insurer and the insured. It is no uncommon thing for business men to place reliance in, and act upon, the verbal agreements and assurances of insurance agents. It sometimes occurs that it is inconvenient or impracticable to do ¡the business, at the time, in any other mode. The law does not require the evidence of the agent’s authority to be made a matter of record, and no man ever requires an insurance agent to show his power of attorney to act as such, before insuring with him. It has always been' supposed, and so all the courts' have held, that when a man assumed to act as an agent for an insurance company, received applications, delivered the policies, and received the premiums (and Eickhoff did all this), that was sufficient prima facie evidence of his agency. And certainly it ought to be, when neither he- nor any. other agent of the company dare take the stand and deny his agency. There is in this country to-day millions of dollars’ worth of property insured, and effectually insured, upon verbal contracts, such as was proved in this case. The doctrine of the ma*365jority of the court does not accord with the understanding and practice of the business community, and puts it in the power of insurance companies to adopt a standard of business integrity much below that which ought to characterize the dealings of reputable business men. The question as to whether such a contract was made is one for the jury to decide upon the evidence, and there was abundant evidence to entitle the plaintiff to go to the jury upon it. It has never been submitted to a jury. It was not passed upon by the circuit court, and has not been argued by counsel. For these reasons the judgment of the circuit court ought to be reversed, and a new trial directed. Not to do so is to deprive the plaintiff, wrongfully, of its constitutional right to have the facts of its case tried by a jury.