This action was commenced by a libel in admiralty and was decided by the District Court favorably to the appellee, the libelant. That interlocutory decree was reviewed by this court and reversed, 71 F. (2d) 415. Upon remand, the cause was tried before a different district judge. The court again found favorably to appellee and referred the cause to a commissioner for the purpose of determining the amount of damage. From that decree this appeal is prosecuted.
The facts out of which this controversy arose are fully set forth in this court’s opinion in the former appeal and we shall not repeat them here. The damage arose by reason of an explosion on appellee’s yacht which appellant had manufactured and sold to appellee, and upon which appellant, by agreement of the parties, had been making some alterations with respect to placing a new stove in the galley.
The general charge of negligence was that the employees of appellant so carelessly and negligently conducted themselves in and about such work as to allow the bilge of the yacht to become filled with gas which caused the explosion. The specific charge of negligence was that appellant’s employees cut off the gas supply pipe leading to the water heater, leaving the end of that pipe open, or insufficiently capped, so as to permit gas to escape, that is to say, they placed a paper disk or gasket in the pipe which they had cut off, which caused the gas to escape into the bilge of the yacht, thus causing the explosion.
Appellant denied the negligence charged, and also urged, though not specially pleaded in the first instance, that appellant’s servant, Koder, in charge of the repairs, had been specifically instructed by appellant’s superintendent, Denne, not to hook up the new stove or turn on the gas until the work had been inspected by Denne, of which fact appellee had knowledge through his agent and servant, Captain Clark; that in violation of that instruction, at the urgent request and command of Captain Clark, Koder, against his will and over his protest, did hook up the stove and turn on the gas before Denne had inspected it, or had been given an opportunity to do so.
The court, upon the first trial, found that such instructions had been given and violated as stated above; that appellant was guilty of negligence in its installation of the gas line with the new stove, and that the explosion occurred in the bilge of the vessel as a result of the escape of gas from the free end of the gas line, which in turn was a result of the negligence of appellant’s said servant in performing that work. Notwithstanding those findings, the court concluded that appellant was liable and decreed accordingly. We thought the conclusion of law and decree were inconsistent with the finding, and accordingly reversed the decree with instructions to permit the answer to be amended as requested and to rehear the cause upon the evidence theretofore given, and such further evidence (especially that of Captain Clark) as either party might wish to adduce.
The case now comes to us on the same evidence introduced at the first trial, plus the deposition of Captain Clark in which he categorically denies every affirmative statement of appellant’s witnesses with respect to appellant’s limitation of its servant’s duties, and disclaims all knowledge, prior to the accident, of such limitation.
There is much surplusage in the findings and conclusions, but among the facts pertinent to the questions raised we note the following: The fuel for the stove and water heater was propane gas stored in liquid form in bottles situated in the stack of the vessel which was located amidships on the upper deck and piped through the bilge into the galley. This gas, which is practically odorless, is one and one-half times heavier than air, and highly explosive in a mixture with air containing from 2.3 to 7 per cent, of the gas. The work of disconnecting the old stove and connecting the new one was performed by appellant’s servant, Koder. Among other things, he cut a square-head cock into the main gas line in the bilge; he connected the new stove by a pipe running from the main gas line through the galley floor; the gas line leading to the hot water heater was left free at the point in the bilge where it had previously been attached, but it remained attached to the main line. Koder attempted to block the flow of gas through the heater line and into the bilge by cutting a piece of cardboard, impregnated with tallow, and inserting it as a gasket between the main line and the heater line. The shapes of these connecting parts were such that the beveled edge of the male part was forced into the gasket, and the amount of force was measured by the strength of the *798workman, magnified by the leverage of the wrench and the power of the screw. The tests of the various joints and connections made by Koder showed no leaks. The pressure test was made by him while the square-head cock cut off the pressure upon the gasket between the two lines. His match test was made while the gas pipes, at the point in the bilge where he worked on them, had only air in them. Immediately after the explosion a jet of flame was seen to issue from the free end of the heater line. It went out immediately when the gas was shut off at the stacks.
The use of paper as a gasket in the manner described by Koder involved danger which should have been apparent to a reasonably prudent man under the circumstances. It was not a standard practice. The explosion took place in the bilge as a result of the escape of the gas from the free end of the gas line, and as a result of Koder’s work. There was no evidence that it could have resulted from any other 'cause. Appellant was guilty of negligence in its installation of the gas line connected with the new stove, and the explosion resulted from such negligence. The gasoline which propelled the yacht, and all piping for it, was aft of a steel watertight bulkhead, fitted skin to skin and keel to deck. There was no explosion of gasoline. No one told Captain Clark that the stove was not to be put in operation until ¿fter inspection tests or approval by Denne, nor did Denne give Koder such instructions.
The conclusions in substance were that appellant was guilty of the negligence as charged, which was the proximate cause of appellee’s damage.
While appeals in admiralty are heard de novo without limitation by the findings and conclusions of the District Court, yet a perusal of this record leaves no doubt in our minds that the facts found were supported by a preponderance of all the evidence.
It is asserted by appellant that the charge of negligence is based solely upon Koder’s use of the gasket, and it contends that such use as a temporary expedient was neither imprudent nor negligent in law. Conceding without admitting the assertion, we think this contention is not sound. It may generally be conceded, as argued by appellant, that negligence is the failure to conform to the standard of the reasonably prudent man, and that the question here presented was whether the plan adopted by Koder was a safe one, and not whether there were safer ones. However, we think a preponderance of the evidence supports the conclusion that the standard employed was neither safe, nor was it the standard of a reasonably prudent mechanic under the circumstances of the situation. Appellant further urges, however, that where, as here, experts disagree on whether the method adopted was safe and efficient, there can be no finding of negligence. We can not accept that statement of the law without qualification. It depends somewhat on the qualifications and credibility of the expert. Under the circumstances here set forth, coupled with the undisputed fact that propane gas is a solvent of tallow, the only charitable conclusion at which we can arrive is that the opinion of appellant’s expert was not based upon any knowledge of propane gas.
It is further contended by appellant that the explosion did not occur in the bilge, and that to so hold would be to base inference upon inference. We think there is no merit in this contention. The gasket and loose end of pipe were in the bilge. All other connections had been tested and found not to leak. Immediately after the explosion, a -flame was observed in the galley, coming from the - bilge, and from the direction of the loose end of pipe in which the gasket was placed. When the gas was turned off in the stack, the flame immediately went out. These were not inferences; they were facts, and, with all the other evidence in the cause, are quite cqnvincing that the explosion occurred in the bilge.
Appellant urges that to warrant a recovery based only on circumstantial evidence, the hypothesis relied upon for recovery must exclude the possibility of other causes of the explosion not attributable to appellant’s fault. Hence, it argues that since other equally probable causes of the explosion did exist, appellee has failed to satisfy this burden of proof. The cases relied upon by appellant in support of this rule are those in which only circumstantial evidence was involved, and no case cited discloses a state of facts analogous to those before us. We are not in accord with appellant’s statement that other equally probable causes of the explosion existed. So far as this record discloses, the only inflammable and explosive substances on board were propane gas and gasoline, and no one contends that the gasoline caused the explosion. There was no evidence to show that the leak of the propane gas *799could have occurred at any other place than at the gasket, except that appellant, without proof, merely suggests that the cook might have opened a burner on the stove without lighting it. We think the evidence in this case preponderantly excludes the possibility of other causes.
Appellant further contends that appellee’s right to recover is defeated by his failure to prove that gas escaped. It is not necessary to repeat what has been- said as a basis for our conclusion that that fact had been proven. From the facts, however, we feel warranted in drawing the inference from the explosion that the gas did leak.
Decree affirmed.