WINTER, Circuit Judge:
These appeals flow from the extensive litigation which resulted from an holocaust on the James River and adjacent shore facilities in the City of Richmond. In the early morning hours of July 5, 1961, Tank Barge #104, with the Tug RAVEN alongside, was moored at the dock of Crown Central Petroleum Company (“Crown”), discharging gasoline into Crown’s storage tanks ashore. A fire erupted which brought about the death of a crew member of the tug— Robert W. Trexler. The tug and the barge were totally destroyed. Crown’s dock and shoreside facilities were severely and extensively damaged. Nearby homes and other property were also damaged or destroyed.
William W. Marshall was the master of the tug. Richard S. Smith was a mate on the tug and also served as a licensed tankerman for the barge which was unmanned. The barge was owned by L & L Towing Corporation (“L & L”) and, at the time of the fire, was operated under charter by Southern Transportation Company, Inc. (“Southern”). The tug RAVEN was owned and operated by Vermillion Towing Corporation (“Vermillion”) .
Myrtle L. G. Trexler, the widow and administratrix of the deceased seaman, sued first, naming as defendants Crown, the tug and barge, Vermillion, L & L, Southern, and the stockholders and chief executive officers of L & L, Southern and Vermillion. The barge owner and charterer, the tug owner, and their chief executive officers and stockholders then instituted proceedings for exoneration from or limitation of liability. Crown filed claims in each proceeding. The tug owner and barge owner then counterclaimed against Crown, asserting that Crown was solely responsible for the widespread damages which had been suffered. In this proceeding a number of individual property owners in the area of the Crown terminal intervened, claiming against the tug and the barge. They also impleaded Crown to claim recovery against it should Crown be held responsible. Trexler’s administratrix also entered the proceeding, claiming against the vessel owners. The tug owner claimed for its hull loss in the barge owner’s proceeding, and the barge owner claimed for its hull loss in the tug owner’s proceeding.
Crown also filed suits against Marshall and Smith, and similar suits were also filed by several of the individual property owners.
All cases were consolidated for trial. Succinctly stated, the district court in its memorandum opinion concluded that the fire and resulting damages were attributable to the tug’s mate Smith’s negligent use of a Coast Guard unapproved flashlight during the discharging operations. The district court held the tug and barge liable in rem, the tug owner and barge charterer liable in personam, and denied to them the right of limitation. It also held Smith liable individually to those parties who had sued him. In the district court’s view the barge’s in rem liability arose from the fact that it was unsea-worthy since it was not equipped with explosion-proof flashlights. The barge owner was exonerated, but the barge charterer was held liable in personam, because it failed to exercise due care to determine if Smith possessed an approved flashlight. The barge charterer was denied limitation because of its knowledge that the barge was not equipped with explosion-proof flashlights. [539]*539Smith was determined to be the mutual employee of the tug owner and barge charterer and, hence, his negligence was attributable to both.
The tug owner was held liable because it had assumed the responsibility of furnishing proper flashlights and had only furnished one. Because the tug owner had privity and knowledge of the negligence and unseaworthiness of the barge, the district court denied it the right of limitation.
Crown was exonerated and awarded judgment against the tug, barge, tug owner, barge charterer and Smith. The individual property owners also obtained judgment against this group. The damages resulting from Trexler’s death were assessed in the amount of $96,388.59 plus interest, and judgment entered against the tug owner under the Jones Act. The barge charterer was also held liable to Trexler under the Virginia Death by Wrongful Act statute, in the amount of $30,000.00. Execution was allowed, however, only in the event that this Court reversed the district court and exonerated the tug, or if the tug owner did not satisfy the full award made under the Jones Act.
The barge, tug, barge charterer, tug owner and Smith have appealed. We conclude that the finding of the district judge that the catastrophe was precipitated by Smith’s use of an unapproved flashlight was clearly erroneous. We conclude that on this record Crown was solely responsible for the tragic events, and that it is liable to Trexler’s estate, the barge owner, the tug owner, and the various owners of property which was damaged or destroyed. We remand the case for the entry of a decree in conformity with our views and further proceedings to terminate the litigation.
- I -
We must first state the general circumstances of what transpired in order to place our analysis of the district court’s findings and our own conclusions in the proper context.
At the time of the fire the barge was moored to the terminal dock, about 110 feet from the storage tanks which were located well up a hillside. Houses were behind the tanks on the other side of a public street. The barge was moored roughly parallel to the shoreline with its aft end north. The tug was moored outboard at the bow end of the barge and facing in the opposite direction. At the time the fire started, the barge was discharging gasoline from its aft tanks (No. 4) into Crown’s shoreside tank No. 19.
The barge was a steel tank barge 194.7 feet long and 39.4 feet wide, having eight cargo compartments, four each on the port and starboard sides. At the aft deck of the barge was a pump with a capacity of 1,100-1,200 gallons per minute (66,-000-72,000 gallons per hour) at 350 RPM’s. The barge had been previously owned by Crown and sold to the barge owner about three months before the fire. At that time it had been equipped with a pump having a pumping capacity of 35,-000 gallons per hour.
The tug, 54.4 feet long and constructed of wood, was arranged with a raised pilot house forward, followed by a cabin with four berths, engine room and galley at the after end of the house. In addition to Marshall (captain) and Smith (mate), the tug’s crew was composed of Grady Carter (cook) and Robert Trexler (deckhand) .
The Crown terminal was in charge of Clarence C. Price, a watchman. Price notified Clinton B. Chadwick, Crown’s assistant plant supervisor, of the barge’s arrival, and Chadwick came to the terminal to attend to the discharge of the barge. Seventeen days after the catastrophe, Chadwick committed suicide, and the only statement of his observations and his part in the events leading up to the fire was the testimony which he gave at a Coast Guard investigation shortly^ after the fire occurred.
The barge was loaded at Norfolk on July 3-4, 1961, with gasoline to be delivered to Crown and others, and each of the two No. 4 tanks contained 52,777 gallons of gasoline. The discharge to [540]*540Crown began at approximately 12:15 A.M. on July 5, 1961. The barge’s cargo hose was connected to the Crown manifold on the dock by the tug crew and checked by Chadwick prior to the commencement of the pumping operation. Leads from this manifold were controlled by valves and connected by pipes to the Crown shore tanks.
It was intended to pump gasoline into Crown’s tanks Nos. 18 and 19. At the beginning of discharge, tank No. 18 could accommodate 56,498 gallons, and tank No. 19, 61,887 gallons. After the fire the valves on tank No. 19 were found to be open and the valves on tank No. 18 closed.
There was evidence that the discharge was carried on at a rate of 66,000 gallons per hour. It continued without interruption or difficulty for approximately one hour until the pumping was shut down by Tankerman Smith upon his becoming aware of the fire. According to Smith, he had been engaged in checking the No. 4 tank on the stern of the barge. The check of the tank was made by using an ordinary metal cased Coast Guard unapproved flashlight. Smith had previously found that the two Coast Guard approved flashlights to which he had access were inoperative, and he had dropped and broken a third unapproved flashlight which he had borrowed from Chadwick. While checking the tank, he heard a shout behind him, turned and saw fire at the forward end of the barge. Smith cut the engine, disengaged the pump, went overboard, swam to the middle of the river, then turned back and swam into the barge and around it onto the Crown terminal shore under the stern of the barge. Although he was covered by soot, Smith was uninjured and suffered neither burns to his person or clothing nor singeing of his hair, eyebrows or eyelashes.
Marshall, Carter and Trexler were in their bunks in the tug cabin immediately before the fire and were alarmed by the extremely pronounced odor of gasoline fumes. They described these fumes as much stronger than those normally encountered during the discharge of a barge. The three men arose from their bunks and started out of the cabin to investigate. Trexler was in the lead, followed by Carter and Marshall. Trex-ler went through the door and stepped over onto the barge deck. As Carter was about to step out of the tug cabin, he looked to his right, i. e., towards the bow of the barge, and saw fire coming from under the forward end of the barge and along the deck of the tug. Carter called to Trexler to return, but Trexler did not do so and continued on across the barge. Carter and Marshall went through the cabin, out onto the port side of the tug and into the water, swimming over to the other bank. As Carter and Marshall were swimming across the river they observed fire all around the stern of the tug and up the hill to the tanks. Carter and Marshall suffered no injuries or any burns to their persons or clothing. Trexler apparently made his way across the barge, onto the dock and partly up the hill of the tank farm. His clothing was almost entirely consumed by the fire, and 95% of the surface of his body was severely burned. He was taken to the hospital where he died a few days later.
A fire alarm was received by the Richmond Fire Department about 1:32 A.M. on July 5, 1961. After the fire department arrived and began to fight the fire, tank No. 19 exploded. The fire extended across East Main Street, in the City of Richmond, and caused considerable damage to and destruction of a number of residences.
- II -
The district judge found that the sole proximate cause of the fire was Smith’s use of the unapproved flashlight. Upon this finding he rested his determination of liability as between the Trexler estate, the barge, the tug, Crown and the property owners, and his resolution of the subsidiary questions of the entitlement of the barge and the tug to limitation of liability and recovery from Crown. We reject this finding.
[541]*541The tug was apparently equipped with two Coast Guard approved flashlights. Both were inoperative at the time that the tug and the barge docked at the Crown facility and the discharge began. Smith, whose duties included general charge of the discharge operations of the tug and barge, borrowed an unapproved flashlight from Chadwick when he found the two approved flashlights would not work. After a few minutes, however, Smith dropped and broke the borrowed flashlight. He then went aboard the tug and obtained another unapproved flashlight from the engine room, and when the fire broke out he was using it to check the contents of the barge’s No. 4 stern tank from which gasoline was being pumped. If Smith’s testimony is to be believed, Smith was leaning down into the No. 4 tank manhole opening with his flashlight when the catastrophe began.
The finding of the district judge was predicated upon his rejection of Smith’s testimony and that of Marshall and Carter who corroborated him, and his acceptance of the testimony of Chadwick and the Crown night watchman Price. The finding was also based on evidence that after the fire the flashlight was found on the deck a few feet from the No. 4 tank manhole and that a fire extinguisher which had been mounted near the opening to the No. 4 starboard tank was also found on the deck, it obviously having been removed from the bracket on which it was held. In the record was the opinion of one expert, contradicted by other experts, that Smith’s use of the flashlight caused the fire, although the district judge did not appear to have placed much reliance on that opinion. An experiment was conducted during trial, which Crown argues showed how Smith could have started the fire without burning or singeing himself.
At the Coast Guard hearing Chadwick testified that after the discharge operation had begun he went to Crown’s office at the northern end of the tank farm, and that when he later came out of the office everything was “quiet, serene, peaceful” on the barge. As he looked in the direction of the barge, suddenly there was fire on the barge, and the fire spread rapidly, arcing with several big flashups. Chadwick was specific that when he first saw the fire it was only on the barge and, specifically, on the “back portion of the barge which was toward me.”
Price, the night watchman, was in the office with Chadwick. He testified that he and Chadwick “jumped” out of the office because they heard someone holler ■ — -not, as Chadwick had testified, that Chadwick left the office in the ordinary course of events. Price continued that when they got out of the office they saw a blaze on the barge 3 to 4 feet high around the whole stern. The fire was somewhere around the center of the barge on the stern. As he and Chadwick ran to the barge, there was an explosion and the fire spread all over the stern and the dock. As they approached the dock Trexler came over it screaming that he knew that he was going to die. The fire arced to the vent of tank No. 18 which exploded. There had been no fire on the ground until tank No. 18 exploded. All told, there were two explosions on the barge and two on the tanks. On cross-examination, it was shown that Price had previously testified during a discovery deposition that the tug was tied to the starboard side of the barge “down around the stern of it some place.” In fact, the tug was tied at the bow end of the barge.
The testimony of Chadwick and Price is mutually contradictory as to how Chadwick and Price happened to observe the fire for the first time. In addition, from Price’s apparent confusion as to the location of the tug vis-á-vis the barge it could be inferred that when he testified that he first saw the fire at the stern of the barge he may have meant the bow. In any event, their testimony is more sharply contradicted by that of Smith, Carter and Marshall.
Smith testified that when he was looking down into the No. 4 starboard tank, located at the stern of the barge, he heard someone holler and turned around. He then looked south, i. e., to the bow of the [542]*542barge, and saw Trexler between the No. 1 tanks and the bow of the tug. Smith saw fire behind Trexler, and Trexler was also on fire.
Carter and Marshall testified that Trexler and they had been conversing while lying in their bunks in the cabin of the tug. The abnormally strong odor of gasoline fumes caused them to arise to investigate. Carter testified that Trexler emerged from the starboardside first and that he, Carter, was so close behind Trexler that he had to catch the screen door to keep it from slamming. As Carter emerged he looked to the stern end of the tug, which was also the direction of the bow of the barge, and saw fire “like a cone twisting, coming from the end.” He stepped back into the cabin, slammed the door, hollered “run to the other side of boat” and jumped overboard. Carter was specific that there was no fire coming from the galley of the tug, but that the fire was on the deck of the barge. He did not know whether the fire came from the rake of the barge or not. He also remembered hollering to Trexler to tell him there was fire. As he closed the door to the cabin the fire came between him and Trexler.
Marshall testified that as Carter began to step on the deck of the tug, flames came past the door. Marshall called to Trexler to come back and to Carter to go overboard. Marshall and Carter both dove into the river from the port side of the tug. Marshall came up from his dive about 15 feet from the tug and he could see that everything on Crown’s hill was on fire and that there was burning fluid around the stern of the tug and the bow of the barge.
When the flashlight was picked up after the fire there remained a light mark on the dock immediately under it. From this fact the district judge inferred that the flashlight had been in this position during the fire. From the additional fact that the fire extinguisher was found on the deck of the barge after the fire, near the bracket in which it usually rested, the district judge concluded that Smith, in spite of his testimony to the contrary, had attempted to use the fire extinguisher to put out the fire since the district judge believed that Smith was the only person who could have had access to the extinguisher. From this inference, the district judge drew the further inference that the fire must have begun where Smith was working, else Smith “would not have even thought of the fire extinguisher in his haste to leave the barge.”
The evidence which we have recited leads us to reject as “clearly erroneous” the district court’s finding that the fire was started by Smith’s use of the unapproved flashlight. In reaching this conclusion we are not unmindful of the fact that the district judge had an opportunity to observe Smith testify and to judge his credibility. Nor do we ignore the fact that Smith was neither an accomplished nor an overly intelligent witness. He contradicted himself in several respects, and he departed from testimony which he had given in a discovery deposition. At the same time, however, at many of the critical points his testimony was corroborated by that of Carter and Marshall, witnesses concerning whose credibility the district judge made no adverse comments. Furthermore, in deciding this case, the district judge placed heavy reliance on the testimony of Chadwick, who never appeared before him and whose testimony is gravely suspect for reasons which will shortly appear. While Rule 52(a), Fed. R.Civ.Proe., expressly provides that findings of fact shall not be set aside unless “clearly erroneous,” and due regard shall be given to the district judge’s opportunity to judge the witnesses’ credibility, we are nevertheless left with “the definite and firm conviction that a mistake has been committed” (United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948)), when the entire record is closely analyzed and proper weight given to the testimony of Chadwick.
Chadwick’s testimony was presented by use of a portion of the proceedings before the Coast Guard in its in-
[543]*543vestigation of the catastrophe. In those proceedings Chadwick was sworn. There was a presiding officer who conducted the proceedings, ruled on the form of questions and ruled on the admissibility of evidence. The object of the inquiry was to fix responsibility for the fire— the object of these proceedings. Full cross-examination on the facts as then known by counsel for the Trexler estate, counsel for Crown and counsel for the barge and tug was permitted. The interests of those present and represented by counsel were substantially the same as those who are parties in these proceedings, but who did not appear in the Coast Guard proceedings. It is true that Chadwick’s cross-examination was carried on before his gauging records and the barge loading records were produced, and that these could have served as a basis for more pointed and specific cross-examination, but the transcript of the Coast Guard proceedings does not disclose that there was any understanding on the record that Chadwick would be subject to further examination when those records were in hand, and Chadwick was advised by the investigating officer only generally “[y]ou are subject to recall.” Correctly, we think, Chadwick’s testimony was admitted at trial. Insul-Wool Insulation Corp. v. Home Insulation, Inc., 176 F.2d 502, 503-504 (10 Cir. 1949); Hertz v. Graham, 23 F.R.D. 17 (S.D.N.Y.1958); McCormick, Law of Evidence, §§ 231-233 (1954); 5 Wigmore, Evidence, §§ 1386-1388 (3d Ed. 1940).
Although admissible, Chadwick’s testimony is entitled to little weight. His testimony was given nine days after the fire and, although all counsel had full opportunity to cross-examine him in the Coast Guard proceeding, it can hardly be supposed that in a catastrophe of this magnitude all of the pertinent facts and issues were then discernible. More importantly, eight days later, on the day after Crown’s barge loading records and Chadwick’s gauging records were furnished to the Coast Guard, Chadwick committed suicide. The record does not show his motive for self-destruction, but, concededly, Chadwick was in charge of the discharging operation and he knew that Trexler had lost his life. While no criminal charges had been lodged, Chadwick must have known that his part in the disaster was one of the subjects of official investigation, if indeed the finger of suspicion was not pointed to him. Under generally accepted rules of evidence flight after the commission of a criminal act constitutes circumstantial evidence of consciousness of guilt and suicide is a form of flight. 2 Wigmore, swpra, § 276; McCormick, supra, § 248. Chadwick’s suicide, therefore, was circumstantial proof of a guilty conscience. It must be deemed a weighty factor in diminishing the credibility of Chadwick’s exculpatory testimony.
Thus, when we consider the evidence which we have summarized in the light of the relatively little weight to be given to Chadwick’s testimony, we can only conclude that the finding that Smith’s use of an unapproved flashlight caused the fire cannot stand. Smith himself denied the fact. There was corroboration from Carter and Marshall. There was physical corroboration from the fact that Smith was neither burned nor singed. The record suggests strongly that burning or singeing of Smith would be avoided only if he dropped the flashlight into an area of inflammable vapor in which he was not standing or if he threw it. From the testimony of the rapidity with which the barge was enveloped in fire, we do not think that the former was possible. Certainly, there is no reason to believe the latter. Like the district judge, we are not impressed with the courtroom demonstration on a micro-scale and under controlled circumstances. . A close reading of the testimony of the expert whose opinion was that Smith started the fire shows that he adopted this thesis only because he accepted the accuracy of Chadwick’s testimony that the fire started where Smith was standing. His opinion turns on the testimony of Chadwick.
Chadwick’s testimony that the fire began at the stern of the barge was weak in itself, as we have stated, and may be [544]*544explained by the testimony of his coworker Price. Price testified that he and Chadwick did not leave the office until after hollering was heard. The hollering must have been that of Trexler because the fire had begun. Once begun, it spread rapidly, according to Carter and Marshall, from the bow to the stern of the barge. Thus, Chadwick’s first observation of fire at the stern of the barge would not necessarily prove that that was where the fire started. Nowhere in his testimony was Smith asked what he did with the flashlight which he said was in hand at the time that he, too, heard Trexler hollering. Its presence after the fire on the deck of the barge near the manhole seems to us to prove nothing, except to negate the suggestion that Smith threw it or dropped it into an area of inflammable vapor in which he was not standing. Nor do we find persuasive the fact that the fire extinguisher was discovered on the deck, removed from its holder. This also seems to us to establish very little. In addition, Smith was not the only person who had access to the fire extinguisher. Trexler was trapped on the deck of the barge with fire around him, and he may have attempted unsuccessfully to use the fire extinguisher to create a path through the flames.
On the record before us the source of ignition of the fire cannot be determined. As unsatisfactory a witness as Smith may have been to the district judge, Chadwick’s testimony simply cannot overcome Smith’s corroborated testimony and be relied upon to place the fault on the barge or tug on the basis of Smith's acts. There is, thus, an insufficient basis to support a finding that the proximate cause of the fire lay in Smith’s use of the unapproved flashlight, or even that his use of the flashlight was a contributing cause of the fire, and the only direct evidence on the subject negatives the possibility that the stove light in the tug’s galley was where the fire began. We, therefore, see no basis on which to impose liability on the tug or the barge, or to deny them recovery from the party who can be determined was at fault.
- Ill -
Fire results from the concurrence of fuel, oxygen and ignition. In a fire like the one with which we are here concerned, it is axiomatic that oxygen is present in abundant supply in the atmosphere, so that sufficient oxygen to make a combustible mixture of oxygen and gasoline may occur without fault as to the source of oxygen. Ignition might well have occurred without fault on the part of the barge, tug, or any other person, i. e., from static electricity, a nail in the shoe of someone striking a steel object, the striking of the steel barge against a metal fixture on the dock, etc. Of direct concern, therefore, is what the record establishes with regard to the availability of fuel. We read the record to establish that Crown, in dereliction of its duty, permitted an unreasonable and dangerous accumulation of gasoline vapors to collect so that when ignition without fault occurred, the holocaust ensued. That the vapors were present and collected is beyond dispute. On the record we may speculate that they resulted from an overflow of Crown’s No. 19 tank, but such a finding is not necessary to impose liability on Crown.
In the ordinary course of the discharge operation which was being carried on, gasoline vapors would be released into the atmosphere. These vapors would emanate principally from the three-inch vent atop the Crown No. 19 tank which was being filled because the gasoline entering this tank would force out the vapors previously contained. To a lesser extent, vapors would also come from a similar vent atop No. 18 tank. The manholes on the port and starboard tanks Nos. 1 and 4 on the barge were open, but only a negligible quantity of vapors would emanate from them because the No. 1 tanks had fire screens over their tops and gasoline was being pumped from the No. 4 tanks. The pumping operation would cause air to flow into the No. 4 tanks to replace the gasoline which was being re[545]*545moved. The displaced vapors from the Crown tanks Nos. 18 and 19 were disbursed into the atmosphere, but the evidence establishes that until a gallon of pure gasoline vapor is diluted with 70 times its volume of air, it will remain combustible and explosive.
Gasoline vapor and a mixture of gasoline vapor and atmosphere are heavier than pure atmosphere alone. In this case the Crown tanks were on a hillside and were higher than the river and the vessels moored at the dock. In addition, July 4-5 was a hot, humid night with little or no discernible movement of air. The terrain, weather conditions and physical properties of gasoline vapors thus combined to cause vapors emanating from Crown’s tanks to go down the hill and to collect in the area in which the barge and tug were moored.
The record admits no finding other than that the accumulation of vapors on the night in question was unusual and dangerous. The fact is that Marshall, Carter and Trexler, all of whom were experienced tankermen and all of whom were accustomed to the odor of petroleum products and not easily alarmed, were moved to investigate the abnormally strong odor of gasoline just before the fire erupted. In addition, numerous residents of East Main Street, who lived at points higher than Nos. 18 and 19 tanks and who had had experience with prior discharge operations, testified that the fumes were the “strongest” and the “heaviest” that they had ever experienced.
The description of how the fire developed, once it had begun, corroborates the conclusion that an abnormally heavy accumulation of combustible vapors collected and that they originated from Crown’s tanks. The fire, once it began, raced uphill, a pattern which, according to expert testimony, would not have been followed unless there was an abnormal concentration of vapors in the area. Although the barge’s tanks exploded early in the fire, various witnesses described the fire as “arcing” to the vents of Crown’s tanks Nos. 18 and 19, and this phenomenon could not be attributed to the spewing of gasoline by the explosion of the barge’s tanks. By the time Marshall and Carter had retreated into the river, surfaced and looked behind them, the whole hill was on fire. There can be no serious dispute that conditions were such that once ignition was supplied there was almost instantaneous conflagration.
At trial Crown’s assistant operations manager for the eastern states testified that Chadwick was the Crown employee responsible for the unloading of the barge and safety of the operations. He stated unequivocally that the operation should have been discontinued if necessary to prevent an unsafe accumulation of gasoline vapors. This testimony recognized the duty, and the legal responsibility for its breach, discussed in Petition of Sinclair Nav. Co., 27 F.2d 606 (S.D.N.Y.1928). In that case fire occurred when the tanker Thompson was discharging gasoline to a barge. The probable source of fuel for the fire was a break in the discharge line causing gasoline to be pumped onto the deck of the Thompson. As a definition of the basis for the determination of liability, the following statement is particularly pertinent because in that case, as here, the precise source of ignition could not be determined:
“No one can say where the highly inflammable and explosive vapors, formed upon exposure of all this gasoline to the air, may have come in contact with sufficient heat to ignite and explode. Some of the witnesses say they first observed the fire on the ship; others, that the first fire they saw was on the barge. But the question is of little importance, because, once the break in the hose line occurred, fire was but the natural and probable consequence of allowing such a quantity of gasoline, in contact with the open air, to flow on the deck and over the side of the ship. One who allows highly inflammable and explosive gases to escape control, under circumstances which are almost certain to result in the destruc[546]*546tion of life and property, can hardly escape liability because those who are injured in a resulting fire are unable to show with precision where and how the fire began. The proximate cause of the fire was the escape of the gasoline, and liability turns upon responsibility for the break in the discharge line, not upon the place where the escaping gasoline first caught fire.” (Emphasis supplied.) 27 F.2d at 607.
As in Sinclair, liability must be imposed on the party who allowed highly inflammable and explosive gases to escape under circumstances which resulted in death and widespread property destruction. That person, on this record, was Crown, and we hold Crown solely responsible.
The record provides a basis from which we can speculate that the abnormal and dangerous concentration of explosive and inflammable gasoline fumes resulted from an overflow of Crown’s No. 19 tank in addition to those escaping from the vents of its Nos. 18 and 19 tanks. The key to this theory of what transpired lies in the testimony of Chadwick.
Chadwick testified that he intended to place the total of 105,554 gallons of gasoline in the barge’s No. 4 tanks into Crown’s Nos. 18 and 19 tanks. Tank No. 18 could accommodate 56,498 gallons and tank No. 19, 61,887 gallons. Chadwick testified that after tank No. 18 was almost filled, he switched the intake of gasoline to No. 19 tank. His testimony in regard to switching stands alone. It is without corroboration, and, as indicated, his testimony is generally suspect. In addition, there is strong circumstantial evidence to controvert him in this vital regard.
All of the witnesses agree that the pumping operation continued for approximately an hour before the fire erupted. In that period of time, the proof established that the barge’s pump was capable of delivering 66,000-72,000 gallons — a quantity in excess of the unused capacity of either tank No. 18 or tank No. 19. The pump on the barge had recently been replaced with the effect of nearly doubling its pumping capacity. Chadwick was familiar with the barge, but he was apparently unaware that the new pump could deliver more than 35,000 gallons per hour. After the fire, the physical evidence established that tank No. 19 was being filled when the fire broke out. If Chadwick’s testimony that he switched tanks during the pumping operation is not believed, manifestly, tank No. 19 could have been caused to overflow.
In addition to Chadwick’s admission that he thought gasoline was being delivered at a much slower rate than was the fact, there was evidence that under the usual, normal procedure for switching the delivery from one tank to another, the services of two employees were required. These employees perform the operation known as “topping” — one examines the level of gasoline from the top of the tank being filled and at the appropriate time signals the other on the ground who then resets the valves at the manifold to direct the gasoline being pumped to another tank. Chadwick admitted that Smith had volunteered his services for “topping” when Chadwick was ready to switch tanks. Both Smith and Price — the only other person who would have assisted Chadwick in topping tank No. 18 had it been filled first —denied that Chadwick called on their help. Chadwick claimed that for some reason that he never explained he departed from the usual procedure on the night in question and switched the pumping from tank No. 18 to tank No. 19 without “topping” tank No. 18, and that he intended to “top” tank No. 18 at a later time. Even more significantly Smith testified that he did not see Chadwick at the valves at the manifold after pumping had begun, as did Price, who was required to punch a time clock mounted near the area where Chadwick must have gone to make the shift if it ever occurred.
Other circumstantial evidence also casts doubt on the veracity of Chadwick’s statement that he made the shift. Marshall and Carter, seconds after the fire broke out, observed the hill covered in [547]*547flames. Trexler, who repeatedly stated his realization that he would die and whose testimony was probably admissible under United Services Automobile Association v. Wharton, 237 F.Supp. 255 (W. D.N.C.1965), said, before his death, that gasoline was running down the hill and that he saw tanks overflowing.
The only corroboration that can be found for Chadwick’s claim that he switched tanks was the testimony that initially the fire arced to the vents of the tanks and that until they exploded, fire was not observed within the catch basin at the base of either tank. From this, it can be inferred that vapors from specific sources and not vapors from more general sources ignited in the first instance and that there was no initial overflow of raw gasoline. This inference may be overcome from the physical fact that raw gasoline does not burn — only the mixture of gasoline fumes and oxygen is combustible — and after the fire, a drainage pipe from the dike surrounding the base of both tanks was inexplicably found to be unplugged so that the dike could have drained as soon as gasoline flowed into it if it were open prior to the time that the fire began. Some corroboration for the theory that the drainage pipe was open and that it did carry off an overflow from a tank is found in the testimony of Marshall and Carter that when they were swimming in the river, they saw fire on the water around the barge.
However intriguing the fai'lure-to-switch theory may be, we make no finding in this regard. We predicate our conclusion of Crown’s liability solely on its failure to have stopped the discharge before a dangerous and combustible accumulation of explosive vapors occurred.
- IV -
The conclusions which we have reached require the exoneration of the barge, the tug, the barge charterer and tug owner. The judgments against Smith should be reversed.
We turn next to the other specific judgments which the district court should enter. Although we have struck down the judgments which Mrs. Trexler obtained in the district court against the barge, tug, barge charterer and tug owner, it does not follow that she should be left without any recovery or be remitted to a new trial. The judgment for $96,-388.59 with interest which Mrs. Trexler obtained against the tug and the tug owner under the Jones Act was rendered, inter alia, on the finding that the deceased was employed by the tug owner. This judgment must be reversed. A like judgment cannot be entered against Crown, because there is no evidence that the deceased was an employee of Crown, and the law is settled that the Jones Act created new rights in injured seamen only against their employers. Cosmopolitan Shipping Co. v. McAllister, 337 U.S. 783, 69 S.Ct. 1317, 93 L.Ed. 1692 (1949); Continental Cas. Co. v. Thorden Line, 186 F.2d 992 (4 Cir. 1951).
Mrs. Trexler also recovered a judgment under the Virginia Death by Wrongful Act statute, 2 Code of Virginia of 1950 § 8-633 et seq., of $30,000.00 with interest against the barge and barge charterer. As to the barge and the barge charterer, this judgment must be reversed because they, we have found, were without fault. But we think that a like judgment should be entered against Crown because (a) it was at fault, (b) the amount of the judgment was fully litigated and recovery in this amount (the maximum amount recoverable under the statute, 2 Code of Virginia of 1950 § 8-636 as amended by Ch. 387, Acts of 1958) was fully supported by the evidence, and (c) it is just and proper to hold Crown liable in this amount without another trial.
Crown’s fault has hereinbefore been shown. On appeal no error has been claimed in the amounts of either judgment obtained by Mrs. Trexler. Our review of the record satisfies us that recovery in the maximum amount permitted by the Virginia statute was amply [548]*548proved, and Crown participated fully in litigating the amount of damages.
Crown argues, however, that we lack jurisdiction to direct the entry of judgment against it because Mrs. Trexler failed to take a cross-appeal from the judgment adverse to her cause of action against Crown when the barge, tug, barge charterer, tug owner and Smith appealed. We disagree.
The judgment under the Virginia statute was obtained in the case (No. 13,216 in this Court) in which Mrs. Trexler sued Crown, the tug and barge, the barge owner and charterer, the tug owner and their chief executive officers and stockholders. Only the tug and the barge appealed, but the appeal by them unquestionably gave this Court jurisdiction over the case. Since Crown and Mrs. Trexler both appeared in this Court as appellees and participated fully in the appeal, we also have jurisdiction over the parties. While ordinarily an appellee— here Mrs. Trexler — cannot secure alteration or modification of a judgment unless he has taken his own appeal, this rule does not state a jurisdictional restriction on federal appellate courts but merely constitutes a “rule of practice.” Langnes v. Green, 282 U.S. 531, 538, 51 S.Ct. 243, 75 L.Ed. 520 (1938); 9 Moore’s, Federal Practice, ¶ 72.05 [4] at 3034-35.
We are persuaded that justice requires departure from this rule of practice in this case. Crown fully litigated the question of its liability and that of the tug and barge, the tug owner and barge charterer before us. It had fair notice and warning and a full opportunity, which was fully exercised, to respond to all of the arguments which could have been advanced by Mrs. Trexler had she taken an appeal. Stated otherwise, the tug, the barge and the other appellants advanced every argument that Mrs. Trexler could have advanced; Crown met all of the arguments; and, on the merits, we have decided against Crown. We accept the admonition of Professor Moore that the traditional federal “rule of practice” should not be permitted to subvert the “just” result which federal appellate courts are obliged to fashion under 28 U.S.C.A. § 2106, and we conclude that the “just” result is that we direct the district court to enter a judgment against Crown under the Virginia Death by Wrongful Act statute in like amount to that which the court granted against the appellants. 9 Moore’s Federal Practice, ¶ 72.05 [4] at 3035 and footnote 9.
As to all property claimants, including the barge and tug, an interlocutory judgment of liability against Crown should be entered and damages assessed promptly.
Crown should pay the costs of these appeals.
The judgments appealed from are reversed, and the case is remanded for further proceedings, in accordance with the directions contained herein.
Reversed and remanded.