MULLIN, J.

Much evidence was given on the trial for the purpose of reforming the application for insurance, so that the answers to the questions in regard to the keeping a watchman in the factory should be made to express the understanding of both parties on that subject. But, inasmuch as the learned justice who tried the cause held that the application needed no reformation; that, by its terms, the insured were not bound to keep a watchman in the factory from twelve o'clock on Saturday night until twelve o'clock on Sunday night; the case is relieved of the questions raised on the trial, as to the competency of the evidence given for the purpose of procuring a reformation of the contract, and whether a case was made by the evidence that would have entitled the plaintiff to that relief.

Before proceeding to ascertain what the construction of the clause of the application under consideration is, it is important to know whether it is a warranty or a representation merely.

The paper which contained the questions and answers in regard to keeping a watchman in the factory is called, in the policy, a survey, and this survey is expressly referred to in the policy and made a part of it.

Angell, in his work on Fire and Marine Insurance, defines a warranty as being a stipulation inserted in writing on the face of the policy, on the literal truth or fulfillment of which the validity of the entire contract depends. The stipulation is considered to be on the face of the policy although it may be written on the margin, transversely, or on a subjoined paper referred to in the policy.

What is said on the subject of a watch in the factory is contained in the survey, filled out by the insured and delivered to the agent of the insurer, and the policy refers to and makes this survey a part of itself. It is, therefore, clearly within the definition of a warranty as laid down by the learned author of the treatise cited, as well as that given by our own courts to that term. ( Jefferson Insurance Company v. Cotheal, 7 Wend. 73; Brown v. Cattaraugus Ins. Co., 18 N.Y. 385; Chase v. Hamilton Ins. Co., 20 id. 52.)

If at the time the survey was made the factory was not in operation, and the statements contained in it as to the watch kept therein is to be considered promissory rather than an affirmative warranty, yet the rights and duties of the parties are not altered. If the promise has not been kept — the condition precedent performed — the insurer is not bound by the policy. (Angell on Insurance, § 145; 2 Duer Ins. 749; 1 Arn. 502.)

The clause of the survey being a warranty, it then becomes important to ascertain its construction, in order to determine whether it has been broken. In construing contracts of insurance, effect must be given to the intention of the parties, as in the construction of all other contracts.

The rule is very clearly stated by Lord ELLENBOROUGH, in Robertson v. French (4 East, 135). The same rule of construction which applies to all other instruments applies equally to this instrument of a policy of insurance, viz: that it is to be construed according to its sense and meaning as collected, in the first place, from the terms used in it, which terms themselves are to be understood in their plain, ordinary and popular sense, unless they have generally in respect to the subject matter, as by the known usage of trade, or the like, acquired a peculiar sense distinct from the popular sense of the same word, or unless the context evidently points out that they must, in the particular instance, and in order to effectuate the immediate intention of the parties to that contract, be understood in some other special and peculiar sense."

It was of the highest importance to the insurer, in order that it might be able intelligibly to decide whether it would assume the risk, or, if it assumed it, to fix the premium to be charged, to know whether a watch was kept in the factory proposed to be insured, at what time such watch was kept, and the means, if any, of determining whether he discharged faithfully his duty. The question "Is there a watchman in the mill during the night?" was a very significant one, and the answer, "There is a watchman nights," was full response to the inquiry. A watch clock being constructed so as to require the watchman to be at it each hour, or his absence would be discovered in the morning, the question whether there was a watch clock was also a very significant one. Their answer was, there was no clock, but the bell was struck every hour, from eight P.M. until it rang for work in the morning, furnished perhaps the next best means of securing watchfulness on the part of the watchman.

The next question to which an answer was required is: "Is the mill left alone at any time after the watchman goes off duty in the morning till he returns to his charge at evening." To which it was answered: "Only at meal times, and on the Sabbath and other days when the mill does not run."

Fires in the factory might be produced in either of four modes: From fire used in the building, from the friction of the machinery, spontaneous combustion, and by an incendiary. There was no danger from fire used in the building, when the building was not occupied, except for a few hours after each day's work closed, and until it was put out. Nor was there any danger from friction, unless the machinery was in motion; nor from incendiaries, unless when there was no person in the mill; but there was constant danger from spontaneous combustion. A watch in the factory during the night afforded a great security against injury from fire from any cause, in the night; and as the danger existed every night in the week, it was important to the insurers to know whether a watch was on hand every night.

It being important to the insurer to know whether a watch was kept every night, and the question put being general whether a watch was kept during the night, the answer that there is a watch nights, must have been understood to apply to every night. No exception being made in the question, and there being an obvious necessity for a watch every night, both parties must have understood the question and answer to apply to every night. If the insured intended to exclude any night, they should have done it clearly and distinctly. It was no more difficult to say that no watch was kept from twelve o'clock Saturday night to twelve o'clock Sunday night, than it was to exclude the Sabbath in the answer to the next question. And the fact that an exception was made in the next answer, is some evidence that none was intended to be made in the first.

It seems to me quite clear, that the answer "there is a watchman nights," is to be understood to mean there was a watchman in the factory every night. But evidence was given on the trial, of a custom of factories in that section of the country not to keep a watch from twelve Saturday night till twelve o'clock Sunday night, and that the answers are to be construed in reference to such custom. "A custom in order to become a part of a contract, must be so far established, and so far know to the parties, that it must be supposed that their contract was made in reference to it. For this purpose the custom must be established, and not casual — uniform, and not varying — general, and not personal, and known to the parties." (2 Parsons on Cont. 53; Dawson v. Kittle, 4 Hill, 107.)

Within the above rule it seems there was no such evidence as would authorize the court to find a custom amongst the factories in the vicinity of the one insured which should be permitted to control the language of the contract in question. The answers to the questions in the survey must be interpreted according to the popular meaning of the language used.

It is insisted by the defendant's counsel that the agent of the insurer was informed that a watchman was not kept in the factory from twelve o'clock Saturday night till twelve o'clock Sunday night, and, therefore, the case is to be considered as if the answers in question were in accordance with the verbal information. But it is well settled that parol evidence cannot be received to control, explain or modify a warranty in an insurance policy. (Angell on Ins. 143; Jennings v. The Chenango Mutual Insurance Co., 2 Denio, 75; Kennedy v. St. Lawrence Mutual Insurance Co., 10 Barb. 285.)

In the 2d Denio the court held that the rule which prevails upon sales of property — that a warranty does not extend to defects which are known to the purchaser — does not apply to warranties contained in policies of insurance; and that parol evidence that the insured truly informed the agent of the insurer who prepared the application as to the situation of the buildings, which differed from the statement in the application, is not admissible. The same propositions were restated in the case in 10 Barb. 285.

The question has been before this court, and I am unable to say what conclusion it has arrived at in reference to it.

In Bidwell v. The Northwestern Ins. Co. ( 19 N.Y. 179), it was held that when a marine policy stated the insurance to be on account of A., loss if any payable to B., and the vessel was warranted by the assured free from all liens, and B. held a mortgage on the vessel subject to two prior mortgages, the insurance was that of A., the owner, on the vessel, and not of B. upon his interest as mortgagee of A.'s equity of redemption; and that the prior mortgage was a breach of the warranty and fatal to the recovery. The answer admitted that the defendant was informed, when the application for insurance was made, of the interest of the mortgagee, and that the loss was made payable to him to secure his interest as mortgagee. The cause was again tried; there was a verdict and judgment for the plaintiff, which was affirmed at the general term, and again brought before this court, and it is reported in 24 N.Y. 302. I do not find that the facts are materially changed, although the learned judge says they were; yet from his summary of them it does not appear that the case was especially changed. It was nevertheless held that the existence of the liens constituted no breach of the warranty. And the dicta in the opinion go the length of utterly sweeping away all the distinctions which have been supposed to exist between warranties in policies of insurance and contracts of sale. If this court is at liberty to consider itself not bound by the decisions on this question, but that it may now establish a rule that it shall deem to be more wise and just, I am not prepared to say that the doctrine of the learned judge, in the case cited, is not the best. But I trust we shall adhere to principles which have been so long settled, and in conformity to which so many contracts have been made, and the abandonment of which will produce great injustice.

The doctrine of the case of Bidwell v. The North Western Insurance Co., last referred to, may stand without interfering with the cases in 2 Denio and 10 Barb., above cited, and I do not think the principles of the case should be extended beyond the facts of that case.

If I am right in supposing that parol evidence cannot be received to vary the warranty in this case, then the evidence given on that point, and received by the judge, should have been rejected.

The next inquiry in order is: Has the warranty been broken? On this subject there is no dispute. It is conceded by the defendant that if the answer to the question as to the watchman cannot be construed as it was construed by the learned judge at the circuit, then it was broken because it is true that no watch was kept from twelve o'clock on Saturday night until twelve o'clock on Sunday night.

The effect of the breach of the warranty is to annul the policy without regard to the materiality of the warranty, or whether the breach had any thing to do in producing the loss.

The effect is very well stated by Marshall, in his work on Insurance, 249. He says: "A warranty being in the nature of a condition precedent, and therefore to be performed by the insured before he can demand performance of the contract on the part of the insurer, it is quite immaterial for what purpose or with what view it is made, or whether the insured had any view at all in making it. But being once inserted in the policy, it becomes a binding condition on the insured, and unless he can show it has been literally performed, he can derive no benefit from the policy. The very meaning of a warranty is to preclude all question whether it has been substantially complied with or not. If it be affirmative, it must be literally true; if promissory, it must be strictly performed. * * * * With respect to the compliance with warranties, there is no latitude, no equity. The only question is: Has the thing warranted taken place or not. If it has not, the insurer is not answerable for any loss, even though it did not happen in consequence of the breach of warranty."

I am unable to perceive any good reason why parties to a contract may not agree that an action for a breach of it shall be brought within a period shorter than that fixed for bringing an action, or that the right of action shall be deemed abandoned. So far from interfering with, it more effectually secures the end sought to be attained by the statute of limitations. This question was directly up in Ames v. New York Union Insurance Co. ( 14 N.Y. Rep. 253), and it was there held that such a condition in a policy of insurance was valid. Had it not been for the waiver of the condition in that case the action would have been barred.

But the plaintiffs meet this defense by alleging, and, as they claim, proving, that it was waived by the company, and the learned judge charged the jury that if the defendant waived this condition the action could be maintained. He further instructed them that if the defendant suggested a postponement until certain attachments were removed, and at the same time were silent in regard to the limitation, that would be a waiver. The circumstances to which the judge alluded in this part of his charge must be the letter of the defendant's secretary, of the 23d June, 1849. That letter was written in reply to a demand by the treasurer of the Glendale Company for pay of the loss, and it informs Mr. Taft, to whom it was written, that the defendant would not enter on any negotiation touching the claim until the garnishee suits mentioned in the letter were removed.

It seems to me that a waiver, to be operative, must be supported by an agreement founded on a valuable consideration or the act relied on as a waiver must be such as to estop a party from insisting on performance of the contract or forfeiture of the condition.

There is not in this case any agreement to waive the condition requiring the suit to be brought within a year; nor is the defendant estopped from insisting on the condition.

If my tenant agrees to pay me rent on a day named, or his lease will be forfeited, and if before the day I agree, for a valuable consideration, to waive the condition, I am bound by the agreement. If, without consideration, I agree that he may pay after the day, and he, by reason thereof, omits to pay at the day, I am estopped from enforcing a forfeiture. But if, without consideration, I assent to a waiver of payment at the day, but before the day withdraw my assent, and insist on performance in such season as to enable him to perform, I am not estopped. Nor was the defendant in this case estopped, even if the above letter, or the negotiation between the officers of the factory and the insurer, could be considered as a waiver of the condition.

But I can find nothing in the evidence which would justify the inference that either party understood at the time there was a waiver. The letter of the 7th August must have removed any impression of the sort from the minds of the officers of the factory company. By that letter they were distinctly informed that the defendant would not recognize any claim against it under the policy, or upon any matter connected therewith. And what is entirely conclusive upon the subject, the factory company, within the year, commenced their action in Connecticut, thus demonstrating, in the clearest possible manner, that they did not then suppose the defendant had waived the condition.

In no aspect of the case am I able to discover any ground on which this action can be maintained. The judgment of the general term must be reversed and a new trial ordered, costs to abide the event.

All the judges concurring, on both grounds, judgment reversed.