On August 1, 1942, the United States Coast Guard acquired for use pending the duration of the war, a 32-foot motorboat, in accordance with Act of Congress, December 16, 1941. Chap. 486, Section 1, 55 Stat. 807, 14 U.S.C.A. § 72. It was thereafter operated in patrolling the Detroit River and adjacent waters, until December 4, 1942, when its engine was removed for overhauling, and the vessel prepared for winter lay up. During its operation, the boat carried a crew of four, with a boatswain’s mate as commanding officer. All these men were paid by the United States Coast Guard, which also caused the boat to be manned, provisioned, navigated, and supplied with all necessaries.
When patrolling operations in 1942 were suspended because of cold weather, the gasoline engine was removed, and the boat thereafter was towed about a mile to Fisher’s Boat Works, to be stored for the winter. After the engine was removed, the gasoline tank located at the rear part of the boat, containing a quantity of gasoline, remained in the boat, with the gasoline feed line of copper tubing % of an inch in diameter and about 10 feet long, attached to the tank. The end of the feed line had been bent upward for about 15 inches where it had been attached to the carburetor before the removal of the engine. The line had a capacity of 4 ounces of gasoline. The valve at the tank where it connected with the fuel line was shut off.
When the boat arrived at the Fisher Boat Works, it was raised by employees of the Fisher Company on chains before being lowered to rest on timbers. On the morning of December 5, 1942, the boat was resting on the chains, not having yet been officially laid up for the winter, when Blaine Bayhurst, one of the members of the crew, went aboard about 9:30 A.M. to remove dishes, foodstuffs, and personal belongings, and to clean up the ship. There were two cabins in the boat. Bayhurst made several trips to the lower cabin for his clothes and during the morning, lighted an alcohol-burning galley stove, which was placed about 4 feet above the floor level of the lower cabin, and fried a steak. He had the flame going for 10 to 15 minutes. While Bayhurst was aboard that morning, Peter Bliznuk, an employee of the Fisher Boat Works, came aboard, opened the small bilge hatch in the floor of the lower cabin, and drained the water which was about 6 inches deep in the bilges, by unscrewing the drain plug. Bliznuk then left, as did Bayhurst. Subsequently, about 1:30 P.M., Bayhurst returned to the boat, and, later, Vernon Qeverdon, the boatswain’s mate in command, joined him about 4:00 P.M. During the afternoon, Bayhurst was in the lower cabin a large part of the time, removing various articles and, for convenience, had hooked open the door at the top of the stairs between the lower cabin and the aft deck. The weather was extremely cold and had been below freezing for several days previously. When the cold weather had set in, the boat was equipped with two kerosene-burning Perfection heaters, one on the upper deck, and one on the lower deck. Regulations provided that these heaters were to be secured to the boat, and they were so secured during patrol operations. However, when the engine was removed from the boat, the heaters were unfastened from the decks and were not thereafter reattached.
*586During the latter part of the afternoon, Cleverdon proceeded to go into the lower cabin. He had an electric lantern which he took with him. Bayhurst followed him down, holding in his hand one of the lighted kerosene-burning heaters. As Bayhurst leaned over to set the heater on the floor of the lower cabin, there was an explosion. Bayhurst was blown against the stairs and the bedding of the bunks thrown on top of him. In the midst of the explosion, Clever-don had rushed up the stairs past Bayhurst. Immediately after the explosion, the lower cabin was covered with flames. The two men, however, were able to escape up the stairway and across the aft deck. The fire destroyed the boat and spread to other ships which were in storage at the Fisher Company premises, causing considerable damage. As a result, the United States of America filed a petition in the district court describing the explosion, claiming that the damage resulting was not caused by any negligence on its part or on the part of any person in charge of the boat or any persons for whom the government might be responsible; and declaring that the damage was occasioned without privity or knowledge on the part of the petitioner. The government further claimed exemption from liability, as charter owner of the boat, as provided in Sections 182-186, inclusive, of Title 46 of the United States Code, 46 U.S.C.A. §§ 182-186. The petition concluded with requests for appraisement of the government’s interest in the Coast Guard boat; for orders directing the issuance of monitions to all persons claiming damages as a result of the explosion, with citations to appear and make proof of their respective claims; for injunctions restraining any proceedings for the recovery of damages arising out of the explosion; that the court either adjudge petitioner not liable for any such damage, or, if the petitioner were adjudged liable, then, that its liability be limited to the value of the interest of the government in the Coast Guard boat after the explosion, and that, to the extent of such value only, decrees be entered in favor of such claimants before a commissioner to be named by the court; and that petitioner be discharged of any further liability.
To this petition, respondent owners whose property was damaged by the fire resulting from the explosion filed an answer and claims, denying that the damages were not caused by the negligence of the petitioner or of those in charge of the motorboat in question. It further denied that petitioner was a chartered owner of the boat; that it was entitled to limit its liability, or to be exempted from liability; and claimed that the losses occurred with the privity and knowledge of the government and that it was liable to the various respondents for damages in the amounts set forth in their answer. The proceedings came on for hearing before the district court, which, at the conclusion of the trial, entered findings of fact and conclusions of law to the effect that the government was the chartered owner of the boat in question at the time of the explosion; that the burden of proof was upon the respondents, here designated as appellants, to establish the negligence of the government; and that they had failed to prove such negligence. The district court, accordingly, entered a decree exonerating the government from all liability as a result of the explosion and fire; and from such decree, the respondents appeal.
This is a proceeding in admiralty. Oregon Railroad & Navigation Co. v. Balfour, 179 U.S. 55, 21 S.Ct. 28, 45 L.Ed. 82; and on appeal, the case is heard de novo. The Cleveco, 6 Cir., 154 F.2d 605. Accordingly, we proceed to review the evidence to ascertain whether there was proof of negligence and whether respondents sustained the burden of establishing their case.
It is the theory of appellants that the explosion was caused by gasoline fumes resulting from gasoline which escaped from its proper place of confinement and formed vapors in the bilges adjacent to the lower cabin. The gasoline vapor which is said to have formed in this fashion along the flooring of the lower cabin is alleged to have come from leaks in the gasoline tank, as well as from gasoline remaining in the exposed feed line, after the removal of the engine.
Appellants submit a considerable body of evidence to substantiate their claims. Four days after the explosion, on December 9, *5871942, the United States Coast Guard constituted a Board of Inquiry to investigate the explosion. The Board had before it the gasoline tank in question, a cylindrical tank about 6 feet long and 20 inches in diameter. To test it for leaks, the Board caused it to be filled with water. As a result, several leaks were discovered.
Captain Morgan B. Mount, Chief Marine Surveyor for the Home Insurance Company, and a man of many years’ experience in the investigation of marine explosions and disasters, was present at the meeting of the Board of Inquiry as an observer and as a representative of the interested underwriters, and had been requested to attend the investigation as an advisor. Immediately after the explosion occurred, he had received a telephone call concerning it, and was soon at the scene of the accident, taking numerous photographs while the boat was in flames. Captain Mount testified to seeing four different leaks in the tank at the time of the inquiry. The day after the explosion, he had made an investigation of the fire, and had examined the woodwork about the tank, which, he testified, showed practically no evidence of fire in that area, nor did the expert for the government attribute the leaks to the heat or fire of the explosion. From Captain Mount’s examination of the tank and the adjacent woodwork, he gave it as his opinion that the tank showed no evidence of heat that would have caused the leaks, and, in his opinion, the leaks had existed before the fire.
It further appeared that when the engine had been removed, the gas line was only disconnected at the carburetor of the engine, and had remained connected to the tank until after the fire. It is also admitted that gasoline was found in the tank even after the fire and explosion. Unless the gasoline, which would naturally flow into the gas line, was blown out when the engine was removed, it would remain in the line. In this case, then, there would have been gasoline left in the line if it had not been blown out; and it could not have been blown out unless the line was disconnected at the tank. It appears from the evidence that it was not so disconnected before the fire. The end of the line at the place where it was disconnected from the engine was bent upward apparently to prevent the line from any dripping or draining of gasoline. The gas line, when full, would hold 4 ounces of gasoline, and this amount, with the proper proportion of air, would result in a dangerous explosive mixture. The fact is undisputed that gasoline vapor will form such a readily ignitable and explosive mixture within a range of 6 parts of gasoline to 94 parts of air, and 1.3 parts of gasoline to 98.7 parts of air; and explosions resulting from such mixtures will exert pressures of 60 to 100 lbs. per square inch. It appears further that, with respect to safety considerations, gasoline has the maximum rating, for hazard, of inflammable liquids. Moreover, since escaping gasoline fumes are likely to accumulate in the bilges, precautions are taken to clean and flush the bilges with water in which there is placed a chemical, trisodium, which removes gasoline that has saturated the framing or planking of the hull of a vessel and eliminates gas vapors that might accumulate in the bilges. But there is evidence to indicate that while the bilge water had been pumped out from time to time, the bilges had not been cleaned and flushed in order to rid the vessel of any fumes that might have accumulated for a period of approximately two months before the explosion.
The government expert witness, William T. Butler, Chief of the Hazard Prevention Section, Merchant Vessel Inspection Division, United States Coast Guard, who testified with respect to explosions on ships and the dangerous characteristics of gasoline fumes, admitted, on cross-examination, that if he had been present at the time the engine was removed from the boat and had the authority to do so, he would have had the gasoline blown out of the gas line as a precautionary measure.
The evidence disclosed that gasoline fumes are heavier than air, and, in a confined space, seek the lowest level; and they will pass through the slightest crack or space in flooring to reach such a level. Four ounces of gasoline which, as has been remarked, was the capacity of the gas line in question, would permeate a compartment of approximately 100 cubic feet, within the limits of the explosive range of gasoline *588vapors. It was testified by the above mentioned government witness that if any gasoline came from the gas line or leaked from the tank, it would have been confined within the area of the engine room compartment which was below the decks of the boat. He further testified that if gasoline had remained exposed to the air in the gas line after the engine had been removed, he would not have been surprised to find the gasoline vapors which escaped therefrom, in the bilges below the floor boards of the lower cabin.
Appellants’ expert witness was uncontradicted in his testimony that leaking gas from the tank and from the gas line would form gasoline vapors in the bilge. The presence of gasoline fumes in bilges is recognized as rendering a vessel very dangerous. In fact, they are considered of such a dangerous nature on ships that have gasoline engines, that the government takes precautions to see that gasoline does not get into the lower sections of a boat. When filling operations are carried on, it is required that the filling pipe plate and plug be tight at the deck and not open at any place underneath between the filling plug on the open deck and the tank itself.
Appellants’ expert witness, Captain Mount, gave his opinion that gasoline fumes which escaped from leaks in the tank and from the gasoline left in the gas line after the engine had been removed, had found their way into the lowest part of the boat in the bilges under the floor of the lower cabin; and that the explosion and fire were caused by these gasoline fumes which had been ignited by the flame of the lighted kerosene heater, when Bayhurst, the crew member, placed it on the floor of the lower cabin, just at the instant of the explosion.
The expert for the government was asked for the first time, on cross-examination, if it was his opinion that the explosion was caused by the presence of gasoline vapors on the Coast Guard boat, and he replied “No.” Counsel for appellants then asked him why he did not think the explosion resulted from gasoline fumes. His first reason was that if the explosion had been so caused, it would be exceedingly doubtful whether the crew member who took the lighted heater into the cabin would have been present at the trial and that he did not think the two men who escaped from the cabin at the time of the explosion could have done so “under their own locomotion.” The fact was that the crew member in question was either knocked unconscious for a time or stunned by the explosion. He testified: “I was thrown back against the stairs. When I came to there was flames all over the cabin.” Furthermore, a door was blown out of the cabin by the force of the explosion. It is a matter of common knowledge that explosions are freakish; and the fact that a man was blown out of a room through a doorway and knocked down and stunned, rather than injured so that he could not appear at the trial, seems a most inconclusive criterion for judging whether an explosion resulted from gasoline fumes, or from some other fumes, or other agencies. The witness further testified that in his opinion, if the explosion had resulted from gasoline fumes, it would have traveled back and caused a similar and almost simultaneous explosion within the area of the engine room compartment. In this regard, the commanding officer of the ship, Cleverdon, testifying for the government, states that he was not sure but thought that there had been two explosions.
The ineffectual nature of testimony, even on the part of an eyewitness — to say nothing of opinion evidence — as to how a gasoline fire travels and where such an explosion commences, is emphasized in The Hygrade No. 18, D.C., 41 F.Supp. 304, 307, where it was said that even if the witnesses in that case were waiting for a fire to start, “with their eyes fixed on the ‘Hygrade’s’ stern, it is extremely doubtful that they could testify with reliable certainty concerning the point of beginning considering the rapidity with which fires of this nature start and spread. The flames of this fire travelled with such speed that it was almost impossible for the human eye, with one look, to detect where the flames started. It will be remembered that all the witnesses * * * took one look and immediately ran for their lives.”
*589The government expert also based his opinion that the explosion did not result from gas fumes on the testimony of Bayhurst and Bliznuk that they did not smell gasoline in the lower cabin on the day of the explosion. It appears, however, as already observed, that such fumes seek the lowest level, and, as appellants point out, it is emphatically stated in a pamphlet entitled “Motorboat Regulations” issued by the United States Coast Guard Headquarters in Washington in March, 1946:
“It must be distinctly borne in mind that all petroleum vapors are heavier than air and consequently accumulate in the lowest part of the spaces containing them, where, being below head level, they are not readily detected by sense of smell and are unsuspected. Such accumulated vapors may lie dormant in the lower part of an engine-room space for a considerable length of time without mishap; however, should a source of vapor ignition be introduced, such as an open flame, a lighted cigarette, an electric spark, etc., a disastrous explosion may result. Hence, it is imperative to prevent the accumulation of explosive mixtures, first by keeping gasoline out of the bilges, (*A half pint of gasoline in the bilge may create a potential explosive power of 5 pounds of dynamite.) and second, by providing adequate means for ventilating such spaces.”
The government argues that there could not have been an accumulation of gasoline fumes in the lower cabin on the day of the explosion, because on that morning, an alcohol stove had been burning there, which would have caused an explosion at that time. But the flame of this stove was 4 feet above the floor, and the flame of the heater, which is alleged to have ignited the gas fumes, was close to the floor. In this respect, the state of facts is comparable to that in Coryell v. Phipps, 5 Cir., 128 F.2d 702, where a person entered an engine room of a vessel operated by a gas engine; struck a match, and proceeded to close several electric switches. As he closed the last switch, a spark emanated from a point approximately a foot to his left, and ignited gasoline fumes, resulting in a disastrous explosion. The fact that the fumes in the present case were not ignited by the alcohol stove in the morning does not make unreasonable the conclusion that they were ignited by the flame of the heater when it was placed on the floor of the cabin, in the afternoon, as is emphasized by the Motorboat Regulations above quoted.
Other grounds given by the government expert for his opinion that it was not an explosion of gasoline fumes are that the fumes would have previously escaped through the bilge drain when the plug was removed, or would have been diffused as a result of the lifting of the deck hatches during the removal of the engine, as well as by air currents resulting from persons going in and out of the cabin, and also because of the movement of the boat when it was towed in the river. Whether, as maintained in the opinion of the government’s expert witness, any gasoline fumes that accumulated in the bilges would have escaped from the vessel through the drain when the bilge plug was removed, or whether such fumes would have been dissipated by the movement of persons in the cabin, or by the motion of the boat while being towed, are, in the very nature of the case, possibilities that are too problematical to consider as persuasive evidence. There is as likely a possibility that accumulated fumes in the bilge under the floor of the lower cabin, in some manner, made their way by the movement or circulation of air, into the lower cabin, through the opening in the floor, after Bliznuk removed the small hatch or floor board in the cabin — and left it open — when he removed the bilge plug in order to drain the water through the opening in the bottom of the vessel.
The government witness also attached great importance to the fact that when the two men escaped from the cabin up the stairs, they ran over a hatch on the deck which was located over the engine room. Because this hatch had not been blown up by the explosion, the witness concluded that *590it could not have been an explosion of gasoline fumes.
Reading the testimony of this expert witness for the government, who was not present at the explosion, who did not see the boat afire, and who' saw none of the physical exhibits such as the tank, the gas line, and other objects, one would almost conclude that no explosion had ever taken place on the boat in question. If it were not an explosion resulting from gasoline fumes, was it, or might it have been, an explosion caused by other fumes or agencies ? In answer to this question, the witness cautiously refused to answer. Finally, at the conclusion of his testimony, which occupies 53 printed pages of the transcript of the record, counsel for appellants, regardless of the fact that he was an adverse witness, gave him every opportunity to express his views, reasonable, speculative, or problematical, about the case.
Because the testimony of this witness is the evidence upon which the government places its chief reliance in this case, it seems to us that the answers of the witness as to his opinion of the cause of the explosion disclose, in themselves, the weight that should be given them. He was asked:
“Q. What was in your opinion the producing cause of this explosion; was it a combination of causes? Just give us your views and where you have learned the facts of the case.
“A. I have listened to the testimony. I have turned it, the whole matter over in my mind. I have drawn upon my experience. I know of several conditions surrounding the operation of small boats, the repair of small boats, the storage of small boats for winter, and I can visualize in my mind a number of things that could have happened on that boat that I don’t know happened, I get into the field of theory entirely. Some of my theories might sound far-fetched, but we have a basis for crediting some of our opinions with reference to these so-called unknown accidents and they are not always compatible with good reason. * * * Now, I don’t know whether I would like to proceed to tell you all of the things that might cause an explosion. I can’t support them.
“Q. I see. * * * You would hesitate to explain your views on that? This is a little bit unusual. * * * I believe the Court, and I believe all of us would like to hear from you about it. * * * I am not really pressing you, but if you would like briefly to give us your thought of what really would cause the explosion, I would really like to give you the opportunity. If you would prefer not to-
“A. (Interposing) : I have qualified myself as an expert. If I am going to be of any value to you or to the Court, you would have to have faith in what I testified to. Now, if I accept this opportunity you give me to elaborate and to give free run to all of my experiences and casualties, and so forth, I might create situations that to your mind would never appear satisfactory and more to the mind of the Court. More concerned, I am, the mind of the Court. I would prefer to stay within the bounds of what I had heard here.
“Had I seen the after-effects of the casualties, had I had an opportunity to examine the boat yard, to see the equipment they have there, to visualize the conditions in the boat yard, I might be able to contribute something very material to the answer here, but, not having had the opportunity to examine, I do not prefer to go on into some causes, some of which would sound practical and some would sound very impractical.”
The foregoing is not helpful in arriving at a just and reasonable determination of the rights of the parties. One thing that does stand out as intelligible in the testimony is the fact that the witness absolutely refused to answer a hypothetical question based upon an assumption that there were leaks in the gas tank before the explosion.' He thought the leaks might have come 'from the pressure exerted on the metal tank, by the explosion, but, on the other hand, refused to consider that the explosion that took place was a gasoline explosion, because there was not enough pressure exerted by it to force up the hatches which were located between the tank and the cabin where the explosion occurred.
Others had seen the leaks, including Captain Mount, the expert witness for appel*591lants. The government expert had not seen the leaks or the tank. Nevertheless, he was certain that there were no leaks in it before the explosion, because of what he considered was “a fact established by normal operations of that vessel,” which fact was not disclosed. Yet immediately after this statement, he volunteered the declaration that if he could have seen the tank after the explosion, he would have been “able to contribute something very material to the answer here,” but that not having had the opportunity for examination, he would prefer not to discuss any possible causes of the explosion. The tank was not produced on the trial; and this brings us to another consideration in the case.
The Board of Inquiry appointed by the Coast Guard to investigate the explosion had before it the gasoline tank, the pipe line, the door that was reported to have been blown out of the cabin, and other physical exhibits. These were last heard of in the possession of the Coast Guard authorities. None of these exhibits was produced in court. Nothing that developed before the Board of Inquiry was submitted in evidence; and no explanation was given why such exhibits were not produced.
It is contended by appellants that the circumstances of the explosion and the proofs in this case established the negligence of appellee, under the doctrine of “res ipsa loquitur.” This doctrine “means that the facts of the occurrence warrant the inference of negligence, not that they compel such an inference; that they furnish circumstantial evidence of negligence where direct evidence of it may be lacking, but it is evidence to be weighed, not necessarily to be accepted as sufficient; that they call for explanation or rebuttal, not necessarily that they require it; that they make a case to be decided by the jury, not that they forestall the verdict.” Sweeney v. Erving, 228 U.S. 233, 240, 33 S.Ct. 416, 418, 57 L.Ed. 815, Ann.Cas.1914D, 905. “When a thing which causes injury, without fault of the injured person, is shown to be under the exclusive control of the defendant, and the injury is such as, in the ordinary course of things, does not occur if the one having such control uses proper care, it affords reasonable evidence, in the absence of an explanation, that the injury arose from the defendant’s want of care.” San Juan Light & Transit Co. v. Requena, 224 U.S. 89, 98, 32 S.Ct. 399, 401, 56 L.Ed. 680. Substantive rights arising from admiralty law are enforced in a manner conforming to admiralty practice. See Garrett v. Moore-McCormack Co., 317 U.S. 239, 63 S.Ct. 246, 87 L.Ed. 239; and in admiralty cases, the doctrine of res ipsa loquitur is applied. Leathern Smith-Putnam Navigation Co. v. Osby, 7 Cir., 79 F.2d 280; Smith v. United States, 5 Cir., 96 F.2d 976; The Rambler, 2 Cir., 290 F. 791; The Great Northern, 9 Cir., 251 F. 826; The Columbia, D.C.N.Y., 25 F.2d 516. See also Rose v. Stephens & Condit Transportation Co., C.C.N.Y., 11 F. 438; Scott v. The London & St. Katherine Docks Co., 3 Hurl. & C. 596.
In a proceeding by an owner of a vessel for exemption from, or limitation of, liability for injuries caused by an explosion occurring in a steering gear room of a vessel, where the evidence disclosed that there was an explosion of leaking gas that would not have ordinarily occurred, it was held that the application of the doctrine of res ipsa loquitur and the consequent finding of liability was justified, in the absence of evidence of the presence of anything that might explode other than such gas. Leathern Smith-Putnam Navigation Co. v. Osby, 7 Cir., 79 F.2d 280.
In the present case, the government argues that to sustain the claims of appellants, the court must build up an inference based on other inferences; and that, before the doctrine of res ipsa loquitur can apply, there must be some direct proof of the cause of the accident. In the Leathern case above cited, a similar claim was made. The court held that such an argument was not tenable where there was proof of an accident; proof that it occurred out of the ordinary course of events; and that it was extraordinary in character. “Proof of the accident and what caused it,” said the court, “need not be by direct evidence or by the testimony of eye witnesses.” In an excellent opinion, the court, speaking through Judge Lindley, continued:
*592“In substantially all cases, the evidence offered in support of the claimant’s case, under this rule, is necessarily indirect or circumstantial, but it is none the less competent. So, here, there was an explosion in a room where there was nothing to cause it other than gases. There were present methane and carbon monoxide. Presence of explosive gases, absence of any other explosives, difficulty with gases in the past and the fact that an explosion could not have occurred, except through some extraordinary cause, create a situation making the accident of such an unusual character as to justify consistently only the finding that the accident was due to negligence of those having possession or control of the things where the injury occurred. This inference rightfully arises, not merely because the accident speaks for itself, but because the facts are such that unless other explanation be given, the only fair and reasonable conclusion is that the accident was due to some omission of duty. This doctrine is directly applicable here.” 79 F.2d page 283.
The origin of the gasoline fires must, in most cases, be established by circumstantial evidence. The New Berne, 4 Cir., 80 F.2d 244. Although there has been no negligence in the original installations of a gasoline tank and equipment in a boat, yet where the evidence shows that there was the presence of gas fumes in an engine room resulting in fumes which became ignited, it is assumed that with the passage of time, some part of the equipment leaked; and the great possibility of damage attendant upon the use of gasoline brings into play the principle that negligence may be based on circumstantial evidence alone. Coryell v. Pilkington, D.C.Fla., 39 F.Supp. 142, 144, affirmed Coryell v. Phipps, 5 Cir., 128 F.2d 702, affirmed 317 U.S. 406, 63 S.Ct. 291, 87 L.Ed. 363. In its opinion in the foregoing case, the district court, after assuming that there were leaks in the equipment, said: “Just when the defective condition of the tanks made them leaky is in doubt. Expert testimony on this is an unsatisfactory character of evidence. . I am satisfied, and find, that there were gasoline fumes present in the engine room, and that their ignition into combustion and fire caused the damage.”
It is contended by the government that in order for the doctrine of res ipsa loquitur to apply, the party charged with negligence must have been in complete control of the instrumentality causing the accident. It is claimed that this requirement was not met for the reason that one of the claimants owned and operated the boat works where the boat was going to be stored, and that its employees had been assisting in laying it up for the winter. See The Great Northern, 9 Cir., 251 F. 826. Where the person injured had complete control of the instrumentality causing his injury, no recovery is allowed. Asprodites v. Standard Fruit & Steamship Co., 5 Cir., 108 F.2d 728. But the boat here in question must be considered to have been under the exclusive control of the Coast Guard at the time of the explosion. It had not yet been laid up for the winter, and the crew member and commanding officer were still on board at the time for that purpose. Certainly, no one else was in control of the boat at the time of the accident. It is enough if the Coast Guard remains as the exclusive controller of all the factors that may have caused the accident. The question is whether the circumstances were such as to justify a finding that the explosion was the result of the government’s negligence. See Jesionowski, Administratrix v. Boston & Maine R. Co., 329 U.S. 452, 67 S. Ct. 401, 91 L.Ed. 416, 169 A.L.R. 947; The Columbia, D.C.N.Y., 25 F.2d 516.
This case calls for application of the doctrine of res ispa loquitur. The proofs show leaks in the gas tank which contained gasoline, even after the fire, as well as gasoline left in the open and exposed gas line — both of which circumstances could have resulted in an accumulation of gasoline fumes in the lower part of the boat. Moreover, there was an accident, extraordinary in character, occurring out of the natural train of events, and such as, in the usual course of things, does not occur if the one, having control of that which causes the injury, uses proper care. Also, there is a complete absence of any explanation for the explosion other than *593one caused by fumes escaping from the leaking tank and the exposed gas line. Further, there is no evidence of regular inspection of the tank or of the bilges. From the considerable body of evidence in the case relating to these conditions and happenings, it is our conclusion that the circumstances of the explosion and the proofs in the case established the negligence of appellee, under the doctrine of res ipsa loquitur.
It is also to be observed that although the tank and other equipment of the boat were in the possession of the government at the time of the trial, they were not produced despite the fact that the government expert testified he would have been able to contribute more in the way of explanation of the explosion and expert opinion evidence if he could have seen the tank. Since it was not produced, the presumption is that its production would have constituted evidence unfavorable to the respondent, and in such a case, the court is justified in concluding that the proof, if offered, instead of rebutting, would sustain the case against the government. Patton-Tully Transportation Co. v. Barrett, 6 Cir., 37 F.2d 516; Henderson v. Richardson Co., 4 Cir., 25 F.2d 225.
We come, then, to the issue whether the government was the chartered owner of the boat at the time of the explosion, for it must show that it was either an owner or a chartered owner to bring its proceedings within the scope of the limitation of the statutes limiting liability. The boat was taken into the Coast Guard pursuant to a written offer and acceptance, which was referred to as a license, and which provided that it might be relinquished by the Coast Guard at any time; but the boat was offered and accepted for the duration of the war. The government obtained exclusive possession and management of the vessel from the owner and became owner pro hac vice. See Hust v. Moore-McCormack Lines, Inc., 328 U.S. 707, 66 S.Ct. 1218, 90 L.Ed. 1534. Accordingly, the government was entitled to maintain these proceedings in admiralty to the same extent as one specifically designated a charter owner. The MV Bull Calf, D.C.Mo., 66 F.Supp. 1019. In Dailey v. Carroll, 2 Cir., 248 F. 466, 467: “The evidence shows the charterer in absolute control and entitled to direct the navigation and employment of the vessel. These are the indicia by which the nature of the charter party is tested and decision reached as to whether it is a demise or not.” The mere fact that the government was not referred to as charterer, in the document in which the boat was offered and accepted by the Coast Guard for the duration of the war, is not of controlling importance, for the government took over its absolute control, subjected it to use as a government vessel, and caused it to be manned, provisioned, navigated, and supplied with all necessaries. We agree with the conclusion of the district court that the government was the chartered owner of the boat within the meaning of the applicable statutory provisions under which the petition to limit liability was filed.
In order that the liability of the government for any negligence in this case be limited, as prayed for in the petition filed in the district court, it must appear that the damages were incurred without the privity or knowledge of the government. See Title 46 U.S.C.A. § 183.
What negligence might the government be deemed to be privy to, or have knowledge of? Certainly, not some isolated act of negligence of some member of the armed forces. In the case before us, the negligence causing the explosion was not due to the conduct of the crew member in taking the kerosene heater into the lower cabin. The regulations, in this respect, provided that the heater be attached to the decks merely to prevent its being upset by the motion of the boat; and the boat was not being operated at the time of the explosion. It was not negligence merely to take the lighted heater into the cabin.
The negligence was the maintenance of the boat in a hazardous condition which resulted from the accumulation of gas fumes which presumably escaped from leaks in the gasoline tank and from the exposed gasoline line connected with the tank. Was the government responsible for this? There is no evidence that the gasoline tank was regularly inspected for leaks, and there ap*594pears evidence that due precaution would require the gas line to be emptied of gasoline when the engine was removed. If the ship became dangerous because of the extremely perilous nature of gasoline and because of the leaks in the tank and the accumulation of fumes from the exposed gas line, someone was negligent. Was there, then, privity or knowledge on the part of the government?
Privity, like knowledge, turns on the facts of particular cases. Coryell v Phipps, 317 U.S. 406, 63 S.Ct. 291, 87 L.Ed. 363. A corporation, as does the government in this case, acts through human beings. The privity of some of those persons must be the privity of the government, else it could always limit its liability. As the Supreme Court said in a similar case involving a corporation: “Hence the search in those cases to see where in the managerial hierarchy the fault lay.” Coryell v. Phipps, supra, 317 U.S. page 411, 63 S.Ct. page 293. “While the cases generally speak of the knowledge of managing officers as being the knowledge of the corporation, the real test is not as to their being officers in a strict sense but as to the largeness of their authority.” In re P. Sanford Ross, Inc., 2 Cir., 204 F.248, 251.
The burden was upon the government to prove that it had no privity or knowledge of negligence, or that there was no privity or knowledge or the means of knowledge of negligence on the part of those to whom it had delegated the duties of commanding, maintaining, and operating the vessel. Coryell v. Phipps, 317 U.S. 406, 63 S.Ct. 291, 87 L.Ed. 363; The Cleveco, 6 Cir., 154 F.2d 605. A leaking gasoline tank on a vessel is a source of the greatest danger. A way in which such a peril could be guarded against and averted would be by regular inspections of the tank. There was no evidence of any such inspections. In the light of the proofs, strong inferences of negligence can be drawn from the leaving of gasoline in the exposed tubing, after the removal of the engine. There is no evidence that anyone in charge of the vessel for the government was interested in taking any precautions in this regard.
Since the burden was upon the appellee to show that it had no privity or knowledge, we have examined the testimony of the officers and crewmen and witnesses sworn on behalf of the government. It consists largely of statements with reference to delegating authority to remove the engine and routine work done by the crew to lay up the vessel for the winter. To sustain the burden of proof that is imposed in cases like this, it may be observed that it is not sufficient to show that the operations and care of the vessel were placed in the hands of men of experience in such matters. We fail to find from the evidence that the government sustained the burden of proving that it had no privity or knowledge, or that there was no privity or knowledge on the part of those to whom it had delegated the duty of commanding and operating the vessel and maintaining it in a safe condition.
In accordance with the foregoing, the decree of the district court is reversed; the petition for limitation of liability is denied; and the case is remanded to the district court for the entry of decrees in favor of appellants in the amount of the damages sustained
TMs MFserpt is from “Motorboat Regulations,” published by the United States Coast Guard, is set forth in appellants’ brief. It is not part of the record on appeal, but its authenticity is not questioned and the statement is not challenged by the government.