This is an appeal in admiralty from a decree of the district court holding the appellant vessel Fort Fetterman solely responsible for damages occasioned on October 5, 1955, by the collapse of a bascule span on a highway bridge owned and operated by appellee, when the Fort Fetterman collided with the bridge which spans the Ashley River and connects the City of Charleston with St. Andrews Parish in the State of South Carolina.
The district judge filed extensive findings of facts and conclusions of law holding that the negligent navigation of the vessel constituted the sole proximate cause of the collision and damages flow*565ing therefrom. The shipowner’s cross-libel seeking rcovery for damages to the vessel was dismissed. See 155 F.Supp. 359. If these findings of fact are supported by credible testimony, it is well settled that they are not to be disturbed unless found to be clearly erroneous. This principle is particularly applicable where, as in this case, the trial judge heard the witnesses and observed their demeanor on the witness stand. Eastern Tar Products Corporation v. Chesapeake Oil Transport Co., 4 Cir., 101 F.2d 30; Hodges v. Standard Oil Co. of New Jersey, 4 Cir., 123 F.2d 362; The Nichiyo Maru, 4 Cir., 89 F.2d 539; Jordan v. Texas Co., 4 Cir., 123 F.2d 614. We are called upon to examine the record to ascertain whether the factual situation reveals manifest error on the part of the district court. Mere expressions of disagreement do not justify a reversal of such findings of fact if credible evidence exists to support the view of the trial judge.
The bridge was constructed during the years 1923-25, pursuant to authorization by the South Carolina Legislature (30 S.C.Stat. 1072), but not in accordance with the plans attached to and made a part of the permit issued on May 23, 1923, by the War Department. There were, for reasons not disclosed by the evidence, material variances between the approved plans and actual construction. For example, the abutments supporting the bascule spans were required to be at a distance of 146 feet apart with no stated provision for tolerance with respect to protrusion of gears, gear rack, machinery, etc., from the outward vertical faces of the abutments. The bascule spans, when fully open, were required to be raised to an angle of 82 degrees, making the extreme extension of the spans in line with the fender systems and a distance of 138 feet apart. Actually, the bridge as constructed permitted the bascule spans to be raised to a maximum angle of only 71 degrees; the extreme extension of the spans, when raised to this maximum degree, were 107.5 feet apart and overhung the fender systems; the gears used to raise the bascule spans extended about 18 inches beyond the outer faces of the abutments; the fendering system on the western side of the opening was approximately 5 feet from the outer face of the abutment, with the bascule span, when fully open, overhanging the fenders by 14.3 feet; and the outer face of the western fender being approximately 3 ½ feet from the outer limits of the protruding gears. These substantial variances must be considered in light of the provisions of 33 U.S.C.A. § 491, 34 Stat. 84, to the effect that where Congress grants authority to construct and maintain a bridge over navigable waters of the United States, “such bridge shall not be built or commenced until the plans and specifications for its construction, * * * have been submitted to the Secretary of the Army and Chief of Engineers for their approval, nor until they shall have approved such plans and specifications * * *; and when the plans for any bridge to be constructed * * *, have been approved by the Chief of Engineers and by the Secretary of the Army it shall not be lawful to deviate from such plans, either before or after completion of the structure, unless the modification of such plans has previously been submitted to and received the approval of the Chief of Engineers and of the Secretary of the Army”.
Jurisdiction with respect to obstructions over navigable streams within the United States is, of course, vested solely in the Government. Public policy dictates that there shall be no material deviation from the plans and specifications approved by the War Department. As was said in United States v. Ingram, 8 Cir., 203 F.2d 91, 95, certiorari denied 345 U.S. 995, 73 S.Ct. 1136, 97 L.Ed 1402:
“Here, there is neither indication nor ambiguity in the statute as a warrant for any judicial dealing with bridge modification, any more than with bridge erection, as a question of obstruction to navigation * * * Deviation from ‘any plans *566* * * approved,’ except upon the basis prescribed in the grant constituted a violation of the statute, whether done during or done after the building of the bridge.”
While we do not hold that an inconsequential deviation constitutes such a violation of the statute as to preclude the offending party from any right of recovery, it is manifestly clear that the deviations with respect to the Ashley River bridge were substantial and constituted a violation of the statute. To hold otherwise would render the statute meaningless. Nor are we impressed with the fact that the violations have continued for a period of approximately thirty years without accident or complaint. What, then, is the effect of such violations as applied to the facts of this case?
The Fort Fetterman is a T-2 tanker, having an overall length of 523' 6", a beam of 68' 2", and a measurement from main deck to keel of 39'. Her gross tonnage is 10,457; her hull is steel; and she is powered by a turbo-electric drive with a single screw. On the afternoon of October 4, 1955, the vessel arrived in the Charleston harbor and anchored in the Cooper River. The cargo of creosote and tar was to be discharged at the Hoppers Company dock, which required passage up the Ashley River through the bridge in question. In accordance with directions from the Charleston Pilots’ Association, a portion of the cargo was pumped into barges to lighten the vessel to a draft not exceeding 15 feet. This operation was concluded at 3:30 A.M. on October 5, at which time the ship drew 14’ 10" fore and aft, resulting in 24' 3" of freeboard, with 6' of her single propeller and 7' of her rudder out of the water.
At 8:20 A.M. the tug Fort Edisto came alongside the vessel, but did not tie up to same. Apparently it was contemplated that the tug would precede the Fort Fetterman on its passage up the Ashley River for the purpose of notifying the bridge to open in time for the vessel’s passage and to assist in docking. A Pilot and an Apprentice Pilot from the Charleston Pilots’ Association boarded the ship at 8:43 A.M., and at 8:55 A.M., the Fort Fetterman weighed anchor and proceeded up the Ashley River toward the bridge with the tug leading the way.
We have no difficulty in agreeing with the district court as to its conclusions of improper navigation on the part of the Pilot and ship’s officers of the appellant vessel. There was a clear 110-foot channel running diagonally through the draw of the bridge, opening at an angle of 20 degrees measured counterclockwise from a perpendicular to the roadway centerline. Slack water tide was undoubtedly preferable to a flood tide or current in negotiating passage through the bridge of a vessel the size of the Fort Fetterman. The ship endeavored to pass through the bridge approximately 45 minutes prior to the recommended time. The wind was from the east at 16 miles per hour, striking the starboard beam of the ship as she approached the bridge. At a distance of approximately 900 yards south of the bridge the vessel passed the Black Can Buoy but, even prior thereto, the Pilot and ship’s personnel were cognizant of the fact that the wind and tide were having some effect on the handling of the Fort Fetterman and concluded that she had to be held up to some extent into the wind. Despite such knowledge, neither the Pilot nor the ship’s officers exerted any particular effort to determine what precise effect, if any, the wind and tide were having on the speed or the set and drift of the vessel, although they well knew that it would be necessary to negotiate a passage of a 68-foot beam vessel through a charted bridge opening of 110 feet. Success in the execution of this passage required care which was obviously absent in this case.
Appellant strenuously urges that the delay in opening the bridge was the cause of the collision. This contention is sufficiently answered by the ship’s personnel who unanimously affirmed that the ship had perfect steerageway and control at 2 knots or 3 knots while she *567was well downstream and “killing time” awaiting the opening of the bridge. The captain conceded that, irrespective of the alleged delay, the vessel was in good position to go through the bridge. The Pilot and third mate testified substantially to the same effect. The evidence of appellant’s own witnsses proves conclusively that the delay, if any, in opening the bridge had no relation to the collision.
At about the time the vessel passed the Black Can Buoy, she veered off to the leeward to get into position to come to her right in order to line up with the draw opening which was on a 20 degree angle to the channel and river. This placed the beam of the ship directly to the east wind, and the force of the current at a 20 degree angle on the starboard quarter, rather than astern as it previously had been. Appoaching the draw from the leeward side contrary to the accepted rule among mariners, the wind and current undoubtedly had the effect of placing the vessel too far to the lee shore. To offset this action, she was given a full right rudder and her speed increased progressively from “Slow Ahead” to “Full Ahead” in an effort to place the vessel to windward and thereby enter the draw opening. This movement brought the ship’s bow to the starboard to such an extent that the bow entered and went through the draw opening, but the stern was still too far to leeward and the vessel “crabbed” into the southwest corner of the fender system and bridge, striking the same with a forward and sideway movement at 9:39 A.M., with the port side of the ship making contact with the west side of the bridge.
The initial contact was about 135 feet aft from the stem of the vessel at a point approximately 40 feet forward of the fashion plate and superstructure. The learned district judge found as a fact that this contact was with the gear rack on the downstream girder of the western bascule, which gear rack was in an illegal position. It is on this finding that the court concluded the “sole proximate cause” of the collision rested upon appellant, as this initial blow broke the gear rack and bent it in an upstream direction. In effect, the lower court held that the resulting damage was inevitable after this initial contact as the downstream girder was caused to fall several degrees so that it dropped in the path of the fashion plate and superstructure. While the evidence on this phase of the case is in sharp conflict, we are willing to accept as a fact that the gear rack was broken and bent in the manner described, but we are unable to agree with the lower court that this initial impact was the sole proximate cause of the collision and resulting damage as the presence of the gears constituted a violation of the permit in protruding beyond the face of the abutment.
With the bascule spans at a maximum elevation of 71 degrees, if the vessel had proceeded through parallel and immediately contiguous to the west fender line, where it admittedly had the right to be, the parties concede that the top of the ship’s superstructure on the port side would have come in contact with the uppermost point of the raised bascule. The district judge held that this fault on the part of the bridge never became an operative factor as the effective damage had already taken place and, furthermore, that such damage, if incurred, would have been speculative and negligible.
It is the legal presumptions and conclusions of the lower court with which we must disagree. It is said by the district judge that any deviation from the approved War Department permit should not be considered unless such deviation was the proximate cause of, or contributed to, the collision and damages. We believe that the factual situation here presented calls for the clear application of the rule as stated in The Pennsylvania, 19 Wall. 125, 86 U.S. 125, 22 L.Ed. 148, to the effect that the burden rests upon the party guilty of a statutory fault to show, not only that probably his fault did not contribute to the disaster, but that it could not have done so. The Pennsylvania rule has been applied in instances of a vessel striking a *568bridge.1 In Dorrington v. City of Detroit, 6 Cir., 223 F. 232, 245, the drawbridge operator failed to open the bridge in violation of the statute and the court, in reference to The Pennsylvania rule, said:
“While it might probably be going too far to say that the failure to open the bridge caused the injury to the vessel, yet the city, having disobeyed the positive statutory mandate * * * should show that its failure could not have contributed to the injury.”
We are not unmindful of the fact that The Pennsylvania rule has been the subject of varying discussions of different degrees of proof required to rebut the presumption that an infringing party is deemed to be at least contributorily negligent. Its application has yielded in many instances to what is commonly referred to as the major-minor fault rule. This latter theory is frequently invoked where there exists a statutory violation which is not deemed material or substantial — a problem which we are not required to meet in the instant case as it is abundantly clear that the statutory violations were material and contributed to the collision.
The lower court, in support of its conclusion, refers to Savannah & New York Transp. Co. v. Klaren Bridge Co., 4 Cir., 252 F. 499. This was an action at law and not on the admiralty side of the court. The jury was left to determine whether minor deviations in the approved plans constituted a statutory fault which was tantamount to a proximate cause of the collision between the vessel and the bridge. We do not believe that our holding in the Klaren Bridge case is controlling in admiralty where the deviations in the plans are substantial.
The structure, of necessity, must be unlawful before The Pennsylvania rule becomes applicable. Authorities such as the Pan Maryland, 5 Cir., 199 F.2d 761, certiorari denied 345 U.S. 909, 73 S.Ct. 649, 97 L.Ed. 1344, and Pacific Spruce Corporation v. City and County of San Francisco, 9 Cir., 72 F.2d 712, are of little assistance because of the absence of such findings.
We are not unmindful of the fact that there may be proper limitations imposed upon The Pennsylvania rule as stated in Webb v. Davis, 4 Cir., 236 F.2d 90, 93:
“The requirements of The Pennsylvania rule must be strictly enforced for the safety of navigation, but it should not be pressed to such an extreme as to justify a division of damages when the accident was undoubtedly due to the negligence of an offending vessel whose actions could not be anticipated.”
To the same effect: Compania De Maderas De Caibarien, S.A. v. Queenston Heights, 5 Cir., 220 F.2d 120, 122, certiorari denied sub nom. Esso Shipping Co. v. Compania De Maderas De Caibarien, S.A., 350 U.S. 824, 76 S.Ct. 52, 100 L.Ed. 736; Seaboard Tug & Barge, Inc., v. Rederi AB/Disa, 1 Cir., 213 F.2d 772; Koch-Ellis Marine Contractors v. Chemical Barge Lines, 5 Cir., 224 F.2d 115; Oriental Trading & Transport Co. v. Gulf Oil Corp., 2 Cir., 173 F.2d 108, certiorari denied 337 U.S. 919, 69 S.Ct. 1162, 93 L.Ed. 1728. Irrespective of these decisions, as well as conflicting views expressed by other courts, we believe that, at best, a critical analysis of the opinions from the United States Supreme Court leads to the conclusion *569that the “reasonable possibility” test is appropriate. Indeed, in 1894 the Supreme Court expressly stated that The Pennsylvania rule refers to “possibility” rather than “probability”. The Martello, 153 U.S. 64, 14 S.Ct. 723, 38 L.Ed. 637. In other cases the court has pointed out that The Pennsylvania rule is the settled rule and should be strictly applied. Richelieu & Ontario Nav. Co. v. Boston Marine Ins. Co., 136 U.S. 408, 10 S.Ct. 934, 34 L.Ed. 398; Belden v. Chase, 150 U.S. 674, 14 S.Ct. 264, 37 L.Ed. 1218; Lie v. San Francisco & Portland S. S. Co., 243 U.S. 291, 37 S.Ct. 270, 61 L.Ed. 726.
The Pennsylvania rule, while not perhaps nullifying the doctrine of proximate cause in admiralty, undoubtedly creates a strong presumption that the material statutory fault is one of the contributing causes of the accident, and places upon the offender the burden of establishing that such violation could not, with “reasonable possibility”, have contributed to the collision. Applying these principles to the facts at hand, we do not believe that the appellee has met the required burden. While appellee suggests that the protruding gear racks were not stressed by appellant at the time of trial in the lower court, the violation existed and was apparent from the drawings. Moreover, were we to disregard the protruding gear racks, we would arrive at a like conclusion because of the deviations from the approved plan. To hold otherwise would require us to conclude that a vessel, passing through the bridge opening at a place it admittedly had the right to be, would sustain only minor damage to its superstructure and, in turn, cause little or no damage to the bridge.
We agree with the district court in its finding that bridge fender systems should not be expected to withstand an impact from a 10,000-ton vessel making 5 knots or more under a “Full Speed Ahead” bell, but we do not feel that the resistance offered by the fendering system could not possibly have caused the ship to deviate to some extent, thus enabling it to avoid striking the protruding gear rack. The force of this argument leads us to the inevitable conclusion that the appellee has not carried the admittedly heavy burden of establishing that its statutory fault could not have, under the “reasonable possibility” test, contributed to the accident and resulting damages. The fact that the approved plans and specifications made no provision for tolerance as to protruding gears; that the gears did protrude a matter of 18 inches from the outer face of the abutment; that the vessel’s superstructure would have collided with the uppermost portion of the raised bascule span had the ship been navigated parallel with the fender system; and that the bridge opening was substantially reduced by reason of the unlawful structure; all point convincingly to appellee’s difficult position in sustaining the burden imposed by reason of these statutory faults. From a review of the record we think it clear that appellee has not shown that the unlawful structure could not, under the “reasonable possibility” test, have contributed to the collision.
Proctors for the respective parties stated, at the time of argument in this court, that the extent of the damage to the bridge exceeded the damage to the vessel. We are not, therefore, called upon to determine the position of appellee with respect to its claim of sovereign immunity. Clearly, the appellee’s claim is subject to setoff or recoupment.
Holding that the district court was in error in concluding from the facts in this case that appellee had met the burden imposed by law, and being of the opinion that the situation herein presented calls for a division of damages, the decree of the district court is
Affirmed in part, Reversed in part, and Remanded.