Lord Sumner:- The Respondents, owners of the steamship "Hatumet," sued the appellants, owners of the paddle-steamer "Punjab", in the High Court of Judicature at Fort William in Bengal, in the Ordinary Original Civil Jurisdiction of that Court, for negligent navigation in the Hooghly on the 27th December 1911. Fletcher, J., who tried the case, decided against them. The Court of Appeal (Sir Lawrence Jenkins, C.J., and Woodroffe, J.) having reversed the judgment of Fletcher, J., the owners of the "Punjab" have brought the present appeal. The case was such as is regularly tried in Courts of Admiralty, and the practical rules of Admiralty Courts are applicable. The only material difference arising from the fact that the action was brought in the Ordinary Original Civil Jurisdiction of the Court is that neither Fletcher, J., nor the Court of Appeal enjoyed, as their Lordships have done, the benefit of the advice of Nautical Assessors. In collision cases, no rule is better; established than this that, where questions of fact alone arise, a Court of Appeal should be most chary of interfering with the decision of trial Judge, who has seen the witnesses and had the opportunity of forming his estimate of them by their demeanour. Only in exceptional cases and for special reasons should a Court which has not had this advantage reverse the judgment of the trial Judge on questions of fact. This rule has long been observed and often laid down. It is unnecessary to cite authorities in which it has been formulated. The Appellate Court, in effect, treated the appeal as a re-hearing of the whole case. Presumably, therefore, their opinion was (although they do not so say in terms) that Fletcher, J., had not rightly addressed himself to the question before him, so that there had been a mis-trial; if they merely differed from Fletcher, J., as to the effect of the evidence, they would not have thought themselves justified in allowing the appeal. Before their Lordships, counsel for the respondents submitted that Fletcher, J., had failed to appreciate the relative importance of the different incidents, and in a question of seamanlike navigation, had allowed his judgment to be influenced, if not determined, by tests inapplicable to two vessels approaching one another so as to involve risk of collision. Before their Lordships, counsel for the respondents submitted that Fletcher, J., had failed to appreciate the relative importance of the different incidents, and in a question of seamanlike navigation, had allowed his judgment to be influenced, if not determined, by tests inapplicable to two vessels approaching one another so as to involve risk of collision. Accordingly, it becomes necessary to consider the facts, not in order to see how their Lordships would have decided at the trial, but in order to see whether Fletcher, J., so misapplied his mind to the incidents proved before him as to warrant the reversal of his judgment on what, after all, are exclusively questions of fact. The scene of the occurrence was the reach of the Hooghly a little below Calcutta opposite the Sibpur College. Here the river runs about east and west, and the deep-water channel is on the south side of the river. On its north side there is an extensive bank called the Sibpur College Sand, where the water is shallow over a considerable distance outwards from the north bank. As the sands rise steeply from the deep channel, their edge, constituting its northern limit, is marked by a line of three buoys : the Upper, Central and Lower College Sand buoys. It was already daylight; the morning was clear and still, and the "Hatumet," which had left the Kidderpore Dock a little earlier, was proceeding down to sea drawing 23 ft. 2 in. forward and 24 ft. 4 in. aft, in charge of a Hooghly pilot named Wheeler. There is no reasonable doubt that some obstruction or appearance of obstruction ahead, caused this pilot to shift her helm and to handle her engines in a series of rapid manoeuvres, the result of which was that she took the College Sand with her stem at a point afterwards fixed by compass bearings as about 400 ft. below the Central College Sand buoy, where she remained fast for some hours and sustained considerable damage. The natural course for an ocean steamer proceeding down this reach would be to keep in mid-channel or a little to the northward of mid-channel, but, on perceiving any vessel coming up-stream, to close somewhat on the line of the buoys under a port helm so as to pass her port to port. The natural course for an ocean steamer proceeding down this reach would be to keep in mid-channel or a little to the northward of mid-channel, but, on perceiving any vessel coming up-stream, to close somewhat on the line of the buoys under a port helm so as to pass her port to port. The plaintiffs' case was that the "Hatumet," coming down mid-channel at half speed, saw the "Punjab" at a considerable distance coming up-river, exchanged port-helm whistle signals with her, edged to the north side of the deep-water channel under a slight port helm, and was about to pass her port to port, when she saw the "Punjab" suddenly take a sheer to the northward, as if acting under a starboard helm (truly a most astonishing and almost suicidal thing to do), so that she began to cross the "Hatumet's" bows at a distance of only 300 ft. The "Hatumet" is about 3,50 feet long. Thereupon the "Hatumet's" engines were put at slow and her helm was hard-astarboarded. Then, after an interval, according to her engine-room log, of one minute, her engines were put full speed ahead for another minute in order to cant her bow to port, but, as she did not cant, they were then put full speed astern. The "Punjab" passed the "Hatumet," apparently, at this juncture, starboard to starboard, but the "Hatumet," with a right-handed screw going astern, took a sheer to starboard about three points off her course, grounded, and remained fast. It may be observed that although the "Punjab" had passed by at a distance of only a few yards, those on board the "Hatumet" did not hail her. They "refrained even from good words." Now it is common ground that the "Punjab," a shallow-draught river paddle-steamer, actually was coming up at this time with two lighters lashed on her port side and one on her starboard. She did sight the "Hatumet" coming down at a considerable distance; she did starboard her helm; and she did pass the "Hatumet" starboard to starboard, and, what is more, the serang, who was directing the navigation at the time, said that, as he understood, when a big ship gave one short blast to a small ship it meant "get out of my way. That way danger lies." Fletcher, J., rejected the "Hatumet's" story, and, as their Lordships think, must have meant that he believed the "Punjab's" serang, though he did not expressly say so. He said in terms that he was unfavourably impressed by the manner of the "Hatumet's" pilot and his assistant, who were called as witnesses, and that he would feel great difficulty in accepting their evidence standing alone. It is true that he attributed to the pilot a suppression of fact, of which he was not guilty; but their Lordships do not think that this was his only or indeed his main reason for doubting his veracity. He certainly was greatly and unfavourably impressed by the fact that, according to the "Hatumet's" case, the "Punjab" was fine on her port bow, only about 300 feet away, when she was seen to be starboarding. To this all the "Hatumet's" witnesses quite unequivocally pledged themselves from the first. On finding that the "Hatumet" was hard and fast aground the pilot sat down and wrote a report then and there, in which he made that statement. The captain signed it as correct. A similar account was entered in the log. The captain and mate were examined on commission in England and adhered to this story, and so did the pilot at the trial. It was so manifest that the "Hatumet's" manoeuvres after perceiving the "Punjab's" alteration of course passed starboard to starboard, could not have been crowded into the short time in which the two ships, after being only about 300 feet apart, would meet and pass, that, at the end of a very long and full trial, the plaintiffs' counsel admitted collision to have been inevitable if the vessels were anything like so near to one another as 300 feet. It is doubtful whether the Appellate Court appreciated that the plaintiffs had been driven to abandon this most important feature of their case. The respondents' very forcible argument was this : If the "Hatumet" gave a port-helm signal it could only have been for some vessel on her port hand; vessels on her starboard hand were out of her way on the shoals, where she could not follow them. The respondents' very forcible argument was this : If the "Hatumet" gave a port-helm signal it could only have been for some vessel on her port hand; vessels on her starboard hand were out of her way on the shoals, where she could not follow them. The rapid and unsuccessful manoeuvres, to which she had resort, must have been taken in view of some vessel navigating so near to her as to involve risk of collision, and that risk could only arise, in the case of vessels about to pass port side to port side, if the approaching vessel suddenly directed her course to port, and appeared to be crossing the "Hatumet's" bows. No steamer but the "Punjab" was in sight, so the thing speaks for itself. To distrust, and still more to reject, the "Hatumet's" case because her pilot misjudged her distance from the "Punjab," showed on the part of Fletcher, J., a lack of nautical judgment. The distance between two vessels approaching one another in a river is peculiarly difficult to estimate with precision. Even at a distance of 1,000 feet (and so their Lordships are advised is the case) such a manoeuvre was unseamanlike and involved risk of collision. This argument, though strong, falls short of being convincing. If there was no other vessel in the neighbourhood of the "Hatumet" except the "Punjab," whose presence, or supposed presence, at close quarters more or less on the port bow could have led to the "Hatumet's" manoeuvres, it might well have been inferred that the "Punjab" was the cause of them and that they would never have been resorted to if the "Punjab" had kept her course so as to pass port side to port side. If is not to be supposed that the "Hatumet" would have been handled in such an urgent manner in order to avoid the imaginary encounter of a nonexistent vessel. Accordingly, though it was enough for the "Punjab" to show that "Hatumet's" witnesses failed to prove their case without herself accounting for the "Hatumet's" manoeuvres, it was evidently important for her to suggest some explanation of those manoeuvres. One suggestion was that the "Hatumet" had approached close to the steep southern edge of the College Sand, and, as seamen say, "smelt the ground," so that she took a sheer to starboard. One suggestion was that the "Hatumet" had approached close to the steep southern edge of the College Sand, and, as seamen say, "smelt the ground," so that she took a sheer to starboard. Their Lordships are advised that in the position and in the depth of water in which the "Hatumet" must have been at the time this could not have happened. The other suggestion is that the manoeuvres are explained by the presence of a country boat somewhere ahead of the "Hatumet." Such a boat there was, and somehow the "Hatumet's" pilot failed to appreciate its presence and position. Eventually he came into collision with it on his starboard bow, though no damage was done. Where this country boat was just previously and how she shaped her course is not known, but, without assuming to decide anything about her movements. their Lordships think that her presence so close to the "Hatumet" at the critical time provides a possible explanation of the "Hatumet's" manoeuvres which is consistent with an acquittal of the "Punjab," if the "Hatumet's" witnesses are once discredited in their story. Their Lordships think, further, that Fletcher, J., was entitled to attach great weight to the fact that from the first the "Hatumet's" witnesses placed the occurrence of the "Punjab's" sudden change of course at a distance from the "Hatumet" of only about 300 feet. The plaintiffs' error was not inconsiderable. The estimate was far from being substantially exact. A distance apart as much as 1,000 feet at the critical time has now been suggested. Whoever was responsible for the signal estimate, it was one to which three persons, all well qualified to judge, made themselves parties within a few hours of the event. They must at least have known whether they felt sure of their estimate or not. It is a grave matter for a litigant in such cases to shift his ground from that taken up in his preliminary act or in such contemporary documents as came into existence in this case and correspond to a preliminary act. When it comes to trebling the distance in question, the position of the defendants is materially prejudiced. It is a grave matter for a litigant in such cases to shift his ground from that taken up in his preliminary act or in such contemporary documents as came into existence in this case and correspond to a preliminary act. When it comes to trebling the distance in question, the position of the defendants is materially prejudiced. Their Lordships are advised that the maneuvers with which the "Punjab" is charged would be dangerous and likely to involve risk of collision, even at a distance of 1,000 feet; but if that distance were accepted, the appellants might have raised, with some prospect of success, the answer that within a distance of 1,000 feet the "Hatumet" ought to have eluded in safety the consequences of the initial negligence, if any, of the "Punjab." As it was, Fletcher, J., considered that he need not decide this issue; it was but lightly touched upon in the Court of Appeal; and the appellants' Counsel, at their Lordships' bar, assuming that the plaintiffs would adhere to their original case, elected not to press it. No further opinion need be expressed on this point beyond saying that, on the course which the "Punjab" admittedly took, so slight an alteration of the "Hatumet's" helm, if promptly made, would have prevented both collision and grounding, that it would not have been right to view such a shifting of their ground as anything but a grave blow to the credit of the "Hatumet's" witnesses. It is on these witnesses that the plaintiffs had to rely for the alleged interchange of port-helm signals. The "Punjab's" serang said that he heard no such signal from the "Hatumet," and Fuller, a so-called independent witness, who did hear her whistle, was not prepared to distinguish between a mere warning blast and a signal meaning "I am directing my course to starboard." So far as he goes, Fuller appears to negative any port-helm signal from the "Hatumet," and he certainly said that the "Punjab" did not whistle at all. A judgment for the defendants might well be affirmed on the simple ground that Fletcher, J., considering the evidence with proper appreciation of its bearing and without misconception as to the law, found as a fact that he could not rely on the "Hatumet's" witnesses sufficiently to hold that she had made out her case. A judgment for the defendants might well be affirmed on the simple ground that Fletcher, J., considering the evidence with proper appreciation of its bearing and without misconception as to the law, found as a fact that he could not rely on the "Hatumet's" witnesses sufficiently to hold that she had made out her case. Certain considerations, however, have been pointed out, which carry the matter a good deal further. A strong attempt was made at the trial to show that, under the Hoogly Rules, it was the duty of both vessels in that part of the river to pass port side to port side, but it was shown eventually that the rule relied on did not apply to the position in question. A practice was proved for shallow-draft river vessels to cross the College Sand to the northward of the line of buoys, and of this practice the pilot of the "Hatumet" was aware. Certain factors governing the positions of the vessels are sufficiently known to make an arithmetical calculation profitable. The "Hatumet's" pilot was well acquainted with the river and well able to estimate distances on it. At a Marine Enquiry into the conduct of the "Punjab's" serang, held about a month after the incident, this pilot gave evidence that, when he first sighted the "Punjab," she was about 1,000 feet below the Lower College Sand buoy, and the "Hatumet" was then 650 to 700 feet above the Upper buoy. To this evidence he eventually elected to adhere at the trial, and this part of his evidence was accepted by Fletcher, J. From this finding their Lordships think that the plaintiffs cannot now escape. The distance between the Upper and the Lower College Sands buoys is accurately known. It is 2,450 feet. The tide was the last of the flood, running up about 1 knot. This is given by the "Hatumet's" assistant pilot. The "Punjab's" serang gave her speed as 5 or 6 knots, and this was not challenged in cross-examination. The "Hatumet's" pilot put her speed somewhat variously, but in substance at about 44 knots. Her engines were at half-speed ahead. The tide was the last of the flood, running up about 1 knot. This is given by the "Hatumet's" assistant pilot. The "Punjab's" serang gave her speed as 5 or 6 knots, and this was not challenged in cross-examination. The "Hatumet's" pilot put her speed somewhat variously, but in substance at about 44 knots. Her engines were at half-speed ahead. Without tying the respondents to unreasonably precise figures, it is easily demonstrable that at these speeds the "Punjab" could not have been still approaching the "Hatumet," where the latter must have been when those manoeuvres began, which shortly afterwards brought her up at the observed point of her stranding. At that time the "Punjab," according to her witnesses, was actually astern of the "Hatumet" in the shoal water to the northward of the buoys. According to Fuller, she could not have been so far upstream, but still she was to the northward of the buoys and not in the position in which the "Hatumet" must needs place her. As a witness, Fuller must be admitted only under very considerable reserve. Of his evidence at the Marine Enquiry, only a brief and obscure report is to be found in the Record. Whatever it meant, it did not agree with what he said at the trial. All the explanation of this disagreement that he could give was that he was called at the enquiry on his subpoena, apparently without having given a proof of his evidence, and, having said what he had to say, was not cross-examined so as to have the opportunity of altering it. The respondents argued that his evidence at the enquiry placed the "Punjab" on the "Hatumet's" pot bow in the act of starboarding across it so as to disappear gradually on her starboard hand till she was shut out altogether. Their Lordships find Fuller's evidence too obsecure to support this view. If this was really what he saw, he must have said much more or much less than he did. His statement is not inconsistent with the "Punjab's" case, but it is insufficient ground for saying that the "Hatumet's'' case ought to have been accepted. Their Lordships find Fuller's evidence too obsecure to support this view. If this was really what he saw, he must have said much more or much less than he did. His statement is not inconsistent with the "Punjab's" case, but it is insufficient ground for saying that the "Hatumet's'' case ought to have been accepted. One point remains: the "Punjab's" natural course would have been to have gained the shoal water on the north side of the reach before passing the Lower College Sand buoy; but the presence of country boats at anchor near that buoy obliged her to keep in the deep-water channel until she was able to round it and begin to cross the sands by passing between it and the Central College sand buoy. Calculations were made by the "Punjab's" superintending engineer at the instance of cross-examining counsel, which were used to show that her turning circle was such as to prevent her from getting across the line between the Lower and the Central buoys till she had advanced within a few feet of the spot where the "Hatumet" grounded. Their Lordships have given careful attention to these calculations, but find them inconclusive. They depend on factors as to the precise position of the country boats and of the "Punjab" relatively to the Lower buoy, both laterally and downstream, at the moment of starboarding, which cannot exactly be ascertained. The sharp zigzag tracks imputed to the "Punjab" seem inexplicable, while apparently she might have made her in-and-out course round the Lower buoy on curves so flat that she would never appear to be on the "Hatumet's" port bow at any material time or even seem to be crossing her course so as to involve risk of collision. Here again the respondents fail to carry their argument far enough to warrant interference with the decision of Fletcher, J, Perhaps the defendants were lucky at the trial, but in the result their Lordships think that the judgment given for them should stand. The learned Judge tried the case at great length, and was assisted by able counsel. He examined the evidence with care and none the less thoroughly because, as he himself observes, he spoke with considerable diffidence on matters of navigation. The learned Judge tried the case at great length, and was assisted by able counsel. He examined the evidence with care and none the less thoroughly because, as he himself observes, he spoke with considerable diffidence on matters of navigation. The judgment of the Court of Appeal does not point out such errors of apprehension on his part or such want of experience as justified them in reversing his decision, and their Lordships have been unable to discover sufficient ground for doing so. They will accordingly humbly advise His Majesty that the appeal should be allowed with costs here and below, and that the judgment for the defendants should be restored. Appeal allowed.