Jenks, J.:
This is an action to foreclose a mechanic’s lien, brought, by a sub-contractor against Koeppel Brothers, contractors, and Glass, the owner of the premisos. The. building which was the -subject of ■ the contract was destroyed by fire before completion'. The trial *459proceeded upon the concession that if Koeppel Brothers could not recover against Glass, the plaintiff could not. Koeppel Brothers by amendment at the trial pleaded : “On said 21st day of April, 1905, while said store or ground floor was in the exclusive possession and under the exclusive control of defendant Glass, a fire occurred therein, through and by the negligence of the defendant Glass and her agents and servants, which was communicated to the rest of said building and said building was destroyed by said fire.”' This, plea presented the issue litigated, on the theory that if the fire was due to the negligence of Glass, the contractors were excused from completion under the contract, and could recover for a part performance thereof. (Vandegrift v. Cowles Engineering Co., 161 N. Y. 435; Dolan v. Rodgers, 149 id. 489 ; Gallagher v. Nichols, 60 id. 438; Logan v. Consolidated Gas Co., 107 App. Div. 392.) The learned Special Term decided th$.t the fire occurred through the negligence of the defendant Glass.
Tlie burden was upon Koeppel Brothers to establish that the fire was attributable to the negligence of Glass. (Stewart v. Stone, 127 N. Y. 506, citing Lamb v. Camden and Amboy R. R. & T. Co., 46 id. 271.) The mere occurrence of the fire does not permit the presumption that it was due to negligence. (Whitworth v. Erie R. Co., 87 N. Y. 413, 420; Stewart v. Stone, supra.) There is no direct evidence as to the cause of the fire. The evidence adduced by Koeppel Brothers was. met by Glass with evidence largely in contradiction. But "without regard"'to it save-so far as it is corroborated somewhat by the evidence given in behalf of the Koeppels, I shall consider the question whether the Koeppels sustainéd the burden of proof.
In the absence of any fact that could permit thé inference of spontaneous combustion or ignition from electric wires or like causes, I think we may assume that the fire was due to some act of man. In Miller v. Steam Navigation Co. (10 N. Y. 437) the court say: “There is no evidence in this case to show how tlie fire, by which the property in' question was destroyed, originated. The presumption, therefore, is that it arose from some act of man. (Angelí on Carriers, § 156.)” The question whether Glass or his workmen caused the fire depends upon circumstantial evidence'— inference from, positive proof of certain other facts. As fire com*460municates fire, there is a possibility that a, man in a room smoking a cigar may be the canse of a fire that afterwards burns therein. .But whether that relation of canse and effect is established depends upon, the other facts -and circumstances, cumulative or exclusive. The court had before it proof of positive character that theretofore there had been furniture moved by Glass into the building, which was in packing paper or excelsior, which practically filled that , floor, and that Siegfred Glass and his workmen were seen “ there ” on the day .of the' fire, smoking cigars or cigarettes; that none of the contractors’ workmen were in the building on the day of the 'fire-save their watchman, and that he did not smoke on that day: On the other hand, there was no suggestion that there was any intentional firing of the building by any one, or that any one would have been benefited by its burning up. A lighted cigar or cigarette in the same building on the s^une floor would not necessarily canse a conflagration, as if it had been brought .in contact- with escaping gas, or inflammable fumes. We know, of course, that lighted cigars and cigarettes are constantly'brought int.o rooms or. buildings and smoked therein, when the room or building contains things inflammable from the -slightest contact with fire.' We know that cigars and cigarettes do not necessarily cast off sparks or floating flecks of. fire. There was no proof that Glass or his workmen while smoking were moving the furniture on that floor shortly before the.fire. There- was no proof that, any one of .these people were in contact with the furniture or the..packing, or in close, proximity to them. There is no proof when Glass or his workmen smoked., There is no proof when they were on the first floor, or even in the building, save that. Lulcoff, the Koeppels’ watchman, testifies - that .Glass’s working people walked up and down on the third floor when lie was eating his dinner. The watchman, Lukoff, had but to look into the ground floor, where he first saw the fire, but he does not testify . to the-presence of Glass or his Workmen on that floor, at any particular time. Proof that none of the Koeppels’ workmen was there on the. day of the fire is not proof that node of the sub-contractors’ workmen (and the record shows that there were at least two sub- ' contractors) was present and working. If we .consider' the testi- . mony adduced by Glass, we find considerable evidence of positive, ' character that' there were other workmen there on this day; that, . *461there were carpenters and plumbers in the building, and that plumbers were seen on the first floor on the morning of that day, who “ had some fire.” Lukoff, the important witness of the Koeppels, testifies that there were some people there, “ but further up ; they were plumbers.” Of course, the court was at liberty to disregard the testimony adduced by Glass, but the testimony as to the presence of plumbers in the building is from both sides, and. practically is not disputed. The use of fire was entirely natural to their calling in constructive work. To my mind the possibility was not, by evidence of concentration or of exclusion, made proof by those who bore the burden. Certainly there was other and further evidence available to them. The plumbers or other.persons must be known. It seems that the plumbers testified before the fire marshal, and the carpenters-were in court. Thus it seems that it was practicable to adduce evidence which -bore on the question of the exclusion of the other persons who were present from any causal relation to the fire. . Starkie on Evidence' (8th Am. ed. p. *865) says: “ What circumstances will amount to proof can never be matter of general definition, the legal testis.the sufficiency of the evidence to satisfy the understanding and conscience of the jury.” In Lopez v. Campbell (163 N. Y. 340) the court say (p. 347): While a material fact -may be established by circumstantial evidence, still, to do so, the circumstances must be such as to fairly and reasonably lead to tlie conclusion sought' to be established, and to fairly and reasonably exclude any other hypothesis.” In Ruppert v. Brooklyn Heights R. R. Co. (154 N. Y. 90) the court say (p. 94): “ In order to prove a fact by circumstances there . should be positive proof of the facts from which the inference or conclusion is to be drawn. The circumstances themselves must be. shown, and not left to rest in conjecture, and when shown it must appear that the inference sought is the only one which can fairly and reasonably be drawn from these facts. (People v. Harris, 136 N. Y. 429.)
I think that there should be a new trial granted.
Hooker, Gaynor and Rich, JJ., concurred; Hirsohberg, P. J., dissented.
Judgment reversed and new trial granted, costs to abide the final award of costs.