The district court and the circuit court of appeals concurred in finding the Victory grossly in fault, and we see no reason for arriving at any other conclusion. In our opinion, the collision was the direct consequence of the Victory's disregard of the rule of the road, and her reckless navigation.

In any aspect of the case, the rule of the road was to keep to the right.

By rule 18 of the regulations prescribed by the act of April 29, 1864, carried forward in section 4233 of the Revised Statutes, 'if two vessels under steam are meeting end on, or nearly end on, so as to involve risk of collision, the helms of both shall be put to port, so that each may pass on the port side of the other.'

The first rule of the supervising inspectors is to the same effect.1

And this was proven to be the usage in the navigation of the Elizabeth river, and known to the master of the Victory.

These vessels were approaching each other in such directions and with such bearings as required them to keep to the right. The distance between Craney Island light and buoy No. 9 was about a mile and one-eighth. When the Plymothian was at buoy No. 9, the Victory was somewhat to the north of, or near, Craney Island. As the Plymothian straightened down the channel, opposite buoy 9, she was near the eastern side, and heading straight down the channel course N. 1/4 OR 1/2 E. She was obliged to swing, on leaving the pier, to round the buoy, but had steadied down as she passed it. The evidence thoroughly established that she was never on the westerly side of the channel. We think the district judge was amply justified in finding as he did that the Plymothian had not, 'in coming out from Lambert's pier, gone over to the west of the channel, near to red buoy No. 22, and had not, after doing so, recrossed the channel to reach its position near buoy 9, as claimed by the Victory's counsel. The tide was not strong enough to force her over there, and it would have been out of her course to have gone there. The testimony is conclusive to that effect.'

The Victory, when in the neighborhood of Craney Island, was either in midchannel, or to the westward of it, and on the Plymothian's port bow. The Victory's witnesses admitted starboarding twice for different schooners, and hard starboarding just before the collision; but both of the lower courts, in accordance with the great weight of evidence, found that the Victory was not prevented by other vessels in the channel, either from going on the westerly side, or from staying there, while the testimony of the Victory's pilot indicated that his object in heading over for the easterly side of the channel was to cut off some of the distance into port, by passing close to buoys Nos. 7 and 9.

Moreover, as immediately after straightening down the Plymothian ported a little, and then hard ported, even if the Victory had been heading at a gradual angle across the channel all the way from Craney Island, the vessels would have been approaching each other from an oblique direction, which would have brought the inspectors' second rule into play, that vessels so situated 'shall pass to the right of each other as if head and head, or nearly so.'

The starboard-hand rule had no application. Although, when the Plymothian started from the pier, her starboard side must necessarily have been down stream as she turned, the vessels were never starboard to starboard after she had rounded buoy 9 and straightened down the channel, and the Victory had passed Craney Island and straightened up S. 1/2 W. Indeed, the rule applicable when two vessels 'are crossing so as to involve risk of collision,' that 'the vessel which has the other on her own starboard side shall keep out of the way,' is ordinarily inapplicable to vessels coming around bends in channels, which may at times bring one vessel on the starboard of the other. It has often been held, as a general rule of navigation, that vessels approaching each other in narrow channels, or where their courses diverge as much as 1 1/2 or 2 points, are bound to keep to port, and pass to the right, whatever the occasional effect of the sinuosities of the channel. Transportation Co. v. Navigation Co., 22 How. 461; Union S. C. Co. v. New York S. S. Co., 24 How, 307; The Vanderbilt, 6 Wall. 225; The Johnson, 9 Wall. 146; The John S. Hasbrouck, 93 U. S. 405; The Berkshire, 33 U. S. App. 531, 540, 21 C. C. A. 169, and 74 Fed. 906.

It is interesting to note that Union Steamship Co. v. New York Steamship Co. was a case of collision in the channel of the Elizabeth river, where the steamship Jamestown, outward bound, took the eastern side of the channel, rounding Lambert's point near the buoy, and proceeded on her course N. 1/4 E. She was struck by the Pennsylvania, coming up, by reason of the Pennsylvania putting her helm to starboard, instead of keeping her proper course, or porting when it became known that the Jamestown was approaching; and it was held that the Pennsylvania was solely to blame.

The principle was embodied in article 21 of the international regulations adopted by the act of March 3, 1885 (23 Stat. 438, c. 354), providing that 'in narrow channels every steamship shall, when it is safe and practicable, keep to that side of the fairway or midchannel, which lies on the starboard side of such ship,' which is article 25 of the regulations adopted August 19, 1890 (26 Stat. 320, c. 802), and put in operation, after some postponements and amendments, in 1897 (29 Stat. 885, 893), and of the act of June 7, 1897 (30 Stat. 96, c. 4).

In The Pekin (1897) App. Cas. 532, articles 21 and 22 of these regulations were considered. Article 21 is given above, and article 22 read as follows: 'Where by the above rules one of two ships is to keep out of the way, the other shall keep her course.'

That was an appeal from the decision of the supreme court for China and Japan, sitting in admiralty, in which the steamship Normandie was alone held to blame for a collision which took place between her and the steamship Pekin in the river Whangpoo on April 3, 1896. The case was thus stated by Sir Francis Jeune, delivering judgment in the privy council: 'At Pootung point the Whangpoo makes a sharp bend towards the south,—returning, indeed, on its course, at something more than a right angle; and to the eastwards of that point the stream is divided by a line of buoys into two navigable channels,—the northern being called the inside, and the southern the outside, channel. The westernmost of these buoys is known as the 'Old Dock Buoy.' The Pekin was proceeding up the inside channel along the line of buoys (that is to say, on the starboard side of that channel), and the Normandie was coming down the river to the southward of midchannel. It is clear that when near the Old Dock buoy the Pekin ported, and that at or about the same time the Normandie starboarded. The Normandie afterwards endeavored to port, but her helm failed to act owing to what is termed the 'Chow Chow' water, which is, it would appear, a well-known area of eddies or whirlpools off Pootung point. The result was that a colision occurred well to the north of the river, and somewhat to the eastward of Pootung point; the stem of the Normandie striking the port bow of the Pekin.

'The evidence is not clear as to the whistles given by the two vessels. The learned chief justice of the supreme court has found that, 'at the same time two blasts of the Normandie's whistle were blown as a signal to the Pekin, those on board the Pekin simultaneously blew one blast of her whistle.' Those on the Pekin dispute the double blast of the Normandie; but their lordships think that, accepting, as they do, the above finding as correct, it may well be that one of the two whistles of the Normandie coincided with the one whistle of the Pekin, and so those on the Pekin heard only one whistle from the Normandie, and believed that only one was given.'

The appeal raised the question of the conduct of both vessels. The Normandie was held manifestly in fault. As to the conduct of the Pekin, two charges were insisted on by counsel; and, after setting them forth, the opinion thus proceeded: 'The first of these charges raises the question, were those two vessels crossing vessels, within the meaning of article 22, and also the further question, whether the Pekin kept her course. The effect of article 22 has been made clear by several authorities. The cases of The Velocity, L. R. 3 P. C. 44, The Ranger, L. R. 4 P. C. 519, and The Oceano, 3 Prob. Div. 60, have explained and illustrated the distinction which exists in the effect of this rule as regards vessels navigating the open sea, and those passing along the winding channels of rivers. The crossing referred to in article 22 is 'crossing so as to involve risk of collision'; and it is obvious that while two vessels, in certain positions and at certain distances, in regard to each other, in the open sea, may be crossing so as to involve risk of collision, it would be completely mistaken to take the same view of two vessels in the same positions and distances in the reaches of a winding river. The reason, of course, is that the vessels must follow, and must be known to intend to follow, the curves of the river bank. But vessels may, no doubt, be crossing vessels, within article 22, in a river. It depends on their presumable courses. If at any time two vessels, not end on, are seen, keeping the courses to be expected with regard to them, respectively, to be likely to arrive at the same point at or nearly at the same moment, they are vessels crossing, so as to involve risk of collision; but they are not so crossing if the course which is reasonably to be attributed to either vessel would keep her clear of the other. The question, therefore, always turns on the reasonable inference to be drawn as to a vessel's future course from her position at a particular moment, and this greatly depends on the nature of the locality where she is at that moment.

'Their lordships have restated these propositions because they appear to them decisive of this part of the present case. They are advised by their assessors, and it appears to them clear, that, having regard to the features of the locality at the time the Pekin ported her helm (that is to say, when she was near the Old Dock buoy), the vessels were not crossing vessels, within the meaning of article 22. It was reasonable for those on the Pekin as, without fault on their part, they did not hear the double blast of the Normandie before they took action with their helm—to assume that the Normandie mandie would take the outside channel, in which case their courses would not cross, or would take the southern side of the inside channel, in which case their courses would indeed cross, but not so as to involve risk of collision.'

We ought to add that in the case before us, even if the steamers had been so far on the starboard of each other as to justify the pilots in considering that they were not meeting 'head and head,' or nearly so, there was no pretense of an agreement to go starboard to starboard, under inspectors' rule 1; nor was this a case for the application of rule 3.

Testing the Victory's conduct by settled rules, she was plainly in fault for not keeping to the right, and in attempting to cross the Plymothian's course, and her speed renders her conduct still more blameworthy. It does not seem to be controverted that at the time of leaving the lighthouse her speed was 5 1/2 knots through the water, and there was a tide force of 2 knots, which would make 7 1/2 knots over the ground. The circuit court of appeals found that, from Craney Island up, her speed through the water was 6 or 7 miles an hour, with a 2-mile tide assisting her, which would make her speed over the bottom 8 or 9 miles. Certainly she must be held to have known that she was approaching the Plymothian so as to involve the risk of collision, and should have slackened her speed, under rule 21, and have stopped and reversed sooner than she did, when she was informed by sight and hearing that her effort to crowd the Plymothian off her rightful course must be unsuccessful.

If she could not port and keep on her own side, she should have reversed at least as early as when her second two-blast signal was blown and not assented to by the Plymothian yet it was the Plymothian that immediately reversed on hearing that signal, and blew three danger signals, while the Victory did not blow her danger signals until after that, and manifestly did not reverse as early as the Plymothian. At the collision the Plymothian's headway had been stopped, but the Victory had such headway on that she threw the Plymothian's bows around to starboard, while her own bows cut through the Plymothian's three decks and stringer plates, a distance of at least 15 inches, and were damaged as far back as 3 feet.

None of the excuses the Victory set up for being on the wrong side of the channel tended to palliate her guilt. The only time at which the Plymothian could have been on the Victory's starboard bow was when the Victory was below Craney Island, and she was bound to govern herself by the bearings of the vessels as they were after she straightened up the channel course from Craney Island south, and then the vessels were port to port. The Victory met a few steamers, but her captain and pilot admitted that they did not force the Victory out of her course; and they were passed near Craney Island, or the lighthouse, or far below the place of collision. So there were two or three schooners in the channel, but both the courts below found that they did not prevent the Victory from doing her duty by porting and keeping to the right, and her presence near them at all was attributable to her having left the western side of the channel.

We need not elaborate, in view of the concurrence of the courts below, and have gone so far into the evidence on this branch of the case because it illustrates the point on which those courts were at variance.

As between these vessels, the fault of the Victory being obvious and inexcusable, the evidence to establish fault on the part of the Plymothian must be clear and convincing, in order to make a case for apportionment. The burden of proof is upon each vessel to establish fault on the part of the other.

The recognized doctrine is thus stated by Mr. Justice Brown in The Umbria, 166 U. S. 404, 409, 17 Sup. Ct. 610: 'Indeed, so gross was the fault of the Umbria in this connection that we should unhesitatingly apply the rule laid down in The City of New York, 147 U. S. 72, 85, 13 Sup. Ct. 211, and The Ludvig Holberg, 157 U. S. 60, 71, 15 Sup. Ct. 477, that any doubts regarding the management of the other vessel, or the contribution of her faults, if any, to the collision, should be resolved in her favor.'

'Collision' was an exception in all the bills of lading, and, laying out of view the 'negligence' and 'flag' clauses, as the damage was occasioned by collision, and within the exception, it rested upon the underwriters in this case to defeat the operation of the exception by proof of such negligence on the part of the Plymothian as would justify a decree against her if sued alone. Clark v. Barnwell, 12 How. 272, 280; Transportation Co. v. Downer, 11 Wall. 129; The City of Hartford and The Unit, 97 U. S. 323, 325; The Ludvig Holberg, 157 U. S. 60, 15 Sup. Ct. 477.

Apparently, it is a hardship for the underwriters on the Plymothian's cargo to be compelled to bear a portion of their own loss; but, if the Plymothian was free from fault, this is merely the result of the limitation of liability acts, the value of the Victory not being sufficient to pay the entire damages sustained.

The circuit court of appeals and the district court arrived at different conclusions in respect of the Plymothian's entire freedom from fault. The district court held that the Plymothian was without blame, while the circuit court of appeals was of opinion that she was not wholly blameless, because she kept her course 'without taking any precaution whatever until too late, and when the pending collision became inevitable.' Whether she may not have been slightly in fault, may be a close question. This is often so when subsequent knowledge of what might have prevented disaster tends to qualify the inquiry as to the prior duty to aver it. But, after all, the question is, as pointed out by Mr. Justice McLean in Williamson v. Barrett, 13 How. 101, whether it was the duty of the master, in the exercise of due care and caution in the management of his vessel, to give a particular order. And, on a careful consideration of the evidence, we think that the Plymothian was not bound to change her course, or to stop and reverse earlier than she did, and these are the only elements of fault imputed to her.

Were the position and course and signals of the Victory such that the Plymothian was bound to change her course, or to stop and reverse sooner than she did?

The Plymothian, in passing the buoy, straightened down the channel course, on the easterly side. The only change she made after straightening down was by porting her helm, which put her closer to the easterly edge of the channel at the time of the collision. She left her pier at 4 o'clock p. m., under half speed, until she rounded the buoy, when the engines were put full speed ahead, 767 yards from the point of collision. The full speed of the Plymothian was 7 knots. The tide was running against her with a force of 2 knots. She had her engines at full speed against the tide perhaps 5 minutes, and both the courts below found her speed over the ground was about 4 miles an hour. The Victory blew two double-blast signals and a three-blast signal before the collision. The evidence of her crew was that the two-blast signals were sounded within a half a minute or a minute of each other, when the steamer was halfway between Craney Island and buoy No. 7, and that the three blasts followed a minute or so later. The Plymothian heard only one two-blast signal and the three-blast signal at that time, as appeared from disinterested evidence on both sides. None of these independent witnesses heard two double blasts and a three-blast signal from the Victory in short succession. All of them who saw or heard the first two blasts testified that that signal was blown when the Victory was to the north of or about Craney Island. And in her petition for limitation of liability the Victory claimed to have blown that signal 'soon after passing Craney Island light,' and placed the Plymothian at that time as 'apparently starting down the river from opposite Lambert's point.' But if she blew two blasts twice, where her captain and pilot said she did, they were blown close together. The Plymothian's evidence showed that she heard but one, and, if the Victory blew two blasts twice within such a short interval as she claims, it would seem that one of them overlapped the whistle from the Plymothian, or it may be that the last two blasts were overlapped by the Plymothian's danger signal.

The evidence largely preponderates that only one signal of two blasts and the signal of three blasts could be heard in the neighborhood of the Plymothian at the time the Victory claims she repeated the two-blast signal and followed it up with three; and we think that the conclusion is sustained by the weight of evidence that the Victory's first two-blast signal was blown before she passed Craney Island, when the vessels were so far apart that the Plymothian would not have been bound to stop if she had heard or seen that signal, and, as there were at least two river steamboats between the vessels at the time, her pilot might well have supposed that the Victory's two blasts were intended for one of them rather than for him. However, the Plymothian did not see or hear that signal, and cannot be held in fault, as contributing to the collision, because she failed to see or hear what would not have been followed by any change of course; and if there were two two-blast signals within a half a minute, then it is fair to conclude, on the evidence, that one of them was overlapped by the whistle of the Plymothian. The Plymothian acted immediately and effectively on hearing the two-blast signal she did hear, and succeeded in stopping her headway; and this when the risk of collision first appeared, which was when the Victory last starboarded. The course of the Victory was, we have said, along midchannel, or to the west of it, and did not involve a risk of collision until she made a change to port under a starboard helm, just before sounding the two-blast signal heard by the Plymothian. Her witnesses admit that they were starting straight up the channel after passing Craney Island light, heading S. 1/2 W. The channel course below Craney Isiand, as shown by the chart, was S. by E.; and in heading S. 1/2 W. her helm must have been ported, and the vessels were port to port. The testimony of the Plymothian's officers and crew was to the effect that the Victory was on their port bow all the way up from below Craney Island to the point where she changed her course just before blowing a two-blast signal, and they are corroborated by independent and disinterested witnesses. The Victory's witnesses testified to starboarding for two or three schooners who were near the point of collision, and this would account for her sheer to port, as observed on the Plymothian, when she took her precautions for safety. Each of these vessels was entitled to presume that the other would act lawfully; would keep to her own side; if temporarily crowded out of her course, would return to it as soon as possible; and that she would pursue the customary track of vessels in the channel regulating her action so as to avoid danger. The Servia, 149 U. S. 144, 13 Sup. Ct. 817; The City of New York, 147 U. S. 72, 13 Sup. Ct. 211; Belden v. Chase, 150 U. S. 674, 14 Sup. Ct. 264.

The rule applicable to them was that each should keep to her own starboard side of the channel. So long as the vessels were port to port, the Plymothian, proceeding at moderate speed, was not bound to stop and reverse on the chance that the other vessel might depart from the rules of navigation. Nor would the Plymothian necessarily have been bound to stop and reverse at once if she had heard a two-blast signal from the Victory at the time when the Victory claims to have blown it. That was, the Victory says, halfway between buoy No. 7 and Craney Island, which would be 600 yards from the buoy, while the Plymothian was 300 yards above the buoy, which would make the distance between the vessels 900 yards, or a half a mile, as the district judge found. The Plymothian was entitled to rely on her repeated single blast to correct the error of the Victory until it was made apparent by a further cross signal, or from a change of heading, that she was persisting in her wrongful course.

The case of The Stephanotis and The Horton, cited to us from the Shipping Gazette and Lloyd's List, of July 26, 1894, was similar in its main features to the one before us. The collision occurred on the river Tyne. On the part of the Stephanotis it appeared that she was bound up on the north side of the channel, at a moderate speed, in a snowstorm, when the Horton was observed about 250 yeards off, also on the north side of the river, in such a position as to make it impracticable for the Stephanotis to pass up between the Horton and the buoys. She thereupon sounded two short blasts and starboarded her helm, afterwards repeating the signal. A short blast was then heard from the Horton, which was seen to be approaching as if under a port helm. The Stephanotis came ahead, helm astarboard, her engines going full speed ahead, as the only means of avoiding a collision. The case for the Horton was that she was keeping to the south of midchannel, going down the river at the rate of about four knots. The Stephanotis was seen coming up the river, half a mile away, bearing right ahead. The Stephanotis thereupon blew one blast, and the Horton responded with one blast, and then ported and steadied her helm. When the vessels were in a position to clear port to port, the Stephanotis blew two short blasts; and the Horton, after replying with one short blast, again slightly ported her helm. The Stephanotis, however, blew two more blasts; and as she was then seen to be under a starboard helm, the Horton's whistle blew a short blast, and her engines were set full speed astern, but the Stephanotis struck the Horton, doing her great damage.

'The advice which we have received from the gentlemen who assist us is precisely that which struck me the moment I heard the circumstances of this case. The advice to us is that there was time and distance for the Stephanotis to have ported after the Horton blew one blast, and that the master of the Horton might reasonably have expected that he would do so. When the master or pilot of the Horton heard the first two whistles, which were the beginning of the other man's starboarding, and when at that moment starboarding could have little effect, he had a right reasonably to suppose that the man who at that moment was starboarding his helm, or was beginning to do so, seeing him there, and being told by him that he was porting, would have ample time and distance, if he did not choose to persist in starboarding, to have put his helm to port, or to have ceased starboarding, and then they would have gone clear. But he persisted in starboarding. Then comes the question, ought the pilot or captain of the Horton to have stopped or reversed, or stopped, even, at that moment when the other was beginning to starboard? The answer is that his beginning to starboard could hardly have been for the Horton, but must have been for some other reason, and that the master and pilot of the Horton might reasonably have supposed that when, if anybody had looked, they must have seen the Horton, and must have known that the Horton was porting, they had a right to suppose that the Stephanotis would stop starboarding immediately, and would have ported her helm. If so, it was not unreasonable and unseamanlike, I think, for the Horton to keep on as she did. By the time that two ship's lengths were passed over by the two vessels approaching each other,—that is, the time of the second blast,—the collision was inevitable. It seems to me that the judgment of the president, and the advice of his trinity masters, was too severe. It was requiring men to do what no man ought to be expected to do under such strange circumstances. Therefore, with great deference, I disagree with the severity of their sentence, and think the Stephanotis ought to be held soley to blame.'

Similar views were expressed by Judge Brown, of the Southern district of New York, in The Florence, 68 Fed. 940.

In respect of both these vessels, the captain was acting as lookout on the bridge, but there was no lookout stationed on the bows. Can it be said that the absence of such a lookout on the Plymothian contributed to the injury? Her captain, pilot, and third officer were all on the bridge, and the view was clear and unobstructed. Would a lookout on the bows have heard or seen more than they did? The bearing of this inquiry is, of course, on the failure of the Plymothian to detect the first alleged two-blast signal blown by the Victory. We have already indicated that the evidence is, to our minds, satisfactory that that two-blast signal was blown when the Victory was below Craney Island lighthouse. If so, the vessels were a mile and one-eighth apart, and the Victory was on the Plymothian's port bow. The Plymothian blew three single-blast signals after that time, and neither of them was dissented from by the Victory until the last one, and the proper maneuvers to avoid the risk then created were promptly taken. But, if it were assumed that the first two-blast signal was blown when the Victory was halfway between Craney Island and buoy No. 7, the vessels were still half a mile apart, with room to correct the proposed erroneous course. And if the Victory blew her second two-blast signal inside of a half minute after the first, as her witnesses testified, and the Plymothian, upon hearing one of the two, immediately acted on it, the failure to hear the other cannot be considered a fault, under the c rcumstances.

We are of opinion that the Plymothian was on her proper course, that she was not bound to anticipate the conduct of the Victory, and that she took all proper precautions as soon as chargeable with notice of risk of collision. The result is that we agree with the district court in the conclusion that the Victory was wholly to blame, and that the Plymothian was free from fault.

Decree of the circuit court of appeals reversed; the costs of that court to be equally divided between the owners of the Victory and the underwriters. Decree of the district court affirmed; costs in this court for preparing and printing the record to be paid by the owners of the Victory; all other costs in this court to be divided equally between the owners of the Victory and the underwriters.