At about dawn in the morning of December 9,1887, the libelant’s two-masted schooner Helen Augusta, came in collision with the brig Havilah in Long Island sound, probably some five miles or thereabouts to the south-west of Faulkner’s Island light. The brig’s stem struck the schooner a little forward of amid-ships, on the starboard side, *876and injured her so that she sank, in about two minutes, under the bows of the brig. The officers and crew were all saved except the wheelsman, who was seen a moment after the collision, but has not since been heard of. The breeze during the night had been moderate, and “ about N. N. E.;” the water was smooth; theweather clear, and good for seeing lights, and both vessels had their regulation lights burning. The schooner, of about 250 tons register, and loaded with coal, was bound for Saybrook, and had, up to the time of the collision, or at least within a few minutes of the collision, been sailing “full and by.” She would sail within -five points of the wind, or a little less, so that her course was supposed to be “about east by north,” and it is so alleged in the libel. On deck were the first mate, the wheelsman, (who was lost.) and, for a short time before the collision, the steward, who had no duties on deck, except to assist when wanted. The only lookout was the mate, who, as he testifies, acted as lookout, walking back and forth on deck. The mate and the áteward testify that the red light of the Havilah was seen about three-fourths of a mile distant, and about two points off the schooner’s starboard bow, and th#t the light continued upon about that bearing, and not more approaching the stem, until the collision; that shortly before the collision the mate hailed the brig to keep off, but no one was seen, and -no change was apparent in the brig’s course; and that the schooner kept her course without change, her sails full and close-hauled, upon the port tack, until the collision. This is, to some extent, confirmed by the m’aster, who was summoned from below, and came on deck a few moments before the collision. The schooner was making about four or five knots per hour. The brig -was about 130 feet long, of 507 tons register, and bound to New York. She had the wind free, and, until shortly before the collision, according to her testimony, was sailing on a course W. \ N. Her master, after passing Faulkner’s Island light, went below. There remained on deck the second mate, who had charge, the wheels-man, and the lookout forward. The mate and the lookout testifj' that they made the red light of the schooner about two and one-half or three points on their port bow, at some considerable distance; that the light continued in about the same direction, or broadened off a little; that, according to orders to report to the captain all lights of vessels, the second mate went aft to report the red light; that he did so, and returned at once, and that, as he was coming out of the companion way, the green light of the schooner was reported by the lookout, with the order “hard-up,” which the second mate immediately repeated to the helmsman, which order was obeyed; and that the collision occurred very shortly afterwards, the brig falling off, as it is claimed, at least one and one-half points to the southward, under her starboard wheel. The master, hearing the order “hard-up,” immediately came on deck, went to the wheel to assist the wheelsman, and remained there until the collision, which occurred very shortly after the rvheel was put “hard-up.” The bowsprit of the brig ran through the after part of the foresail, between the schooner’s fore and mainmast. The brig’s starboard braces were then hauled back, so as to back the ship away; but, before this could *877take effect, the schooner sank, and her mainspring stay, catching upon the brig’s jib-boom, broke it a couple of feet outside of the cap.

As the brig was sailing free and the schooner close-hauled, it was the legal duty of the brig to keep away from the schooner, and the brig must, therefore, be held responsible for not doing so, when, as in this case, there are no extraneous circumstances to excuse her, unless it appears that the collision was brought about by some fault of the schooner. The fault of the schooner alleged by the respondents is that she did not keep her course. To make this a valid defense the change of course must be something more than a change made in extremis, when collision is inevitable, and that situation is brought about wholly by the other vessel. In such cases a change of course, for the purpose of attempting to escape a collision immediately impending, though the maneuver may he a mistaken one, is not a fault. Steam-Ship Co. v. Rumball, 21 How. 384; Bentley v. Coyne, 4 Wall. 509; The Fairbanks, 9 Wall. 420; The City of Paris, 1 Ben. 174; The Jupiter, Id. 586, 537.

The respondents contend that such a change in her course to the northward is proved—First, by the change in the light exhibited by the schooner from red to green shortly before the collision; and, second, by the angle of collision, which, instead of being but one and one-half points, the difference in their courses, was, as they contend, from four to seven points, proving, as is claimed, in connection with the brig’s alleged keeping off two points, a luff by the schooner of from lour to seven points to the northward; and they contend that there would have been no collision except for that change. The libelant’s witnesses testily that there was no change in the schooner’s course, and that the position of the vessels and the course of the schooner were such that the schooner’s red light was at no time exhibited to the brig. If the light first seen by the lookout of the brig, and reported as red, ivas in reality the schooner’s red light, the contradictions between the libelant’s witnesses and the respondents’ cannot, so Jar as I see, be reconciled; one or the other must he rejected. If, as the respondents’ witnesses say, the schooner’s red light was visible during several minutes, and bore, up to within a minute of collision, from two and one-half to three points, as they allege, off the brig’s port bow, during all that time tile brig’s red light must have been seen on the schooner’s port bow only. It could not have been seen on her starboard bow until brought there by her own luffing within half a minute of the collision. But the first mate and the sLeward of the schooner both say that they saw the brig’s rod light for several minutes, during the interval, about two points off the schooner’s starboard how; that the first mate, by the use of glasses, while the brig was a very considerable distance off, saw that she was a square-rigged vessel; and that they stood at the companion-way watching her on their starboard how with her royals set, and that they had no doubt that she would keep away to the southward, and, until she came near, had no apprehension of collision. If it was the schooner’s red light that the lookout and the second mate of the brig saw, the numerous details in the testimony of the first mate and steward of the schooner on this subject must be a fabrication. Again, a *878change by the schooner of from four to seven points to the northward is not a change possible to be attributed to any carelessness or inattention of the wheelsman. A change of two points, a minute before the collision, would have set her sails shaking, which would have been immediately observed and corrected. The mate and steward testify that repeated orders were given to the wheelsman to keep her full and by. If there was any change of from four to seven points, it must have been a change made deliberately and upon the mate’s order. A change of three points might have been made at the last moment by the wheelsman under the terror of immediate collision. No order to change the course before that could have been given by the mate, except for the purpose of tacking. There was no occasion for tacking at that time, and the place was improper. He says he did not order any change, and it is incredible, if the brig’s red light had been seen on the schooner’s port bow, that she should have tacked deliberately across the brig’s course when the brig was very near. To support the respondents’ contention, in order to avoid that violent improbability, it would be necessary to find that the brig’s red light was not seen at all on the schooner until after she had changed her course by. tacking, and had thereby brought it on her starboard bow, within a half minute ofthe collision; and that the mate deliberately tacked without looking to see what was close to him to windward; and that all the material parts of his narrative and the steward’s are fabrications.

A'careful examination of the testimony of the mate and steward of the schooner, as well as their appearance and deportment as witnesses, impress me with a conviction of their 'truthfulness and capacity. • There are numerous details in their testimony from which I find it difficult to believe fabrication, recklessness, or incompetency. To account for the contradictions, thatis, I think, the least probable alternative. Their story is in itself entirely consistent, natural, and probable, although some doubt and difficulty remain as to the angle of collision. As respects the light exhibited by the schooner to the brig, it is opposed by two witnesses only from the brig, and as regards their testimony, the following circumstances must be observed:

(1) If the red light seen was the red light of this schooner, and if it was seen during an interval of from two to four minutes before the green light was recognized, and if it.bore from two and one-half to three points on the brig’s port bow, as her witnesses say it bore, the collision could not possibly have occurred; for the schooner, from that bearing, could not have reached the brig by any possible course that would have kept her red light exposed until shortly- before the collision, as the brig’s witnesses testify it was. To have enabled the schooner to reach the place of collision, showing her red light until within a minute or half minute of the collision, she could not have borne over one point off the brig’s port bow, and scarcety so much, even, as a point. While some errors in the estimate of bearings should not be held incompatible with the credibility of the witnesses, the -difference between one point and “2£ or 3 points” in the bearing of a light of which the witnesses are bound to take careful notiee, is too great to be ascribed to mere error in *879observation by competent seamen. So great a mistake must necessarily detract from the reliance to be put upon their general testimony. Both the second mate and the lookout of the brig do state repeatedly that the red light which they saw. bore two and one-half to three points on the port bow, increasing somewhat, as they thought, to about three points. The captain says it was reported three points off the port bow. If this was seen, as the respondents claim, some three minutes before collision, the necessary conclusion from this discrepancy, and tlio impossibility of the schooner’s reaching the'place of collision, would be that the red light seen was not the schooner’s, but some other red light. There are other circumstances which lend to diminish confidence in the account of these two witnesses for the brig;

(2) The lookout’s testimony is peculiar;

“Question. Bid you see the light of this schooner? Answer. I seen a red light. Q. And where away was that light? A. It was about two points and a half, perhaps throe points, on the port bow. Q. How faraway do you think she was? A. How far away? I couldn’t tell exactly the distance; it was too dark. 1 only see the light; perhaps three quarters of a mile, or something like that. Q. Did \on make any report? A. Yes, sir. Q. What? A. Ire-ported to the second mate, ‘ Red light on the port bow.’ Q. Where was the second mate when you reported to Mm? A. He was walking amid-ships. Q. When you reported to the second mate, what did Ms do? A. lie came right forward, and came right on the forecastle head. Q. Where you were? A. Yes, sir. Q. When he was on the forecastle head what did Ms do? A. He told me what light it was,—he didn’t kiiow exactly what light it was. Ho said something about a light, and said, ‘ I am going to report to the captain. Captain told me to report hi m the lights.’ Q. Bid you point outto him this red light that you saw? A. Yes; I did. Q. After he wont to report to tlio captain, did yoii see any change in that light? A. I seen it coining towards us, and at once suddenly it was gone. I was walking across, you know. Q. Walking across deck? A. Yes; when I looked at it again I couldn’t see it; it got out of my sight; I didn’t see it any more. I kept looking at it for a while, and at once 1 seo a green light. Q. Where was the groen light? A. She was on the port bow. Q. How much on the port bow ? A. I don’t know exactly; it was a little further than the red light; some over three points. Q. When you saw that light, what did you do? A. I reported it. Q. What did you say? A. • Green light on the port bow.’ Q. Did you say anything else? A. I didn’t say anything else. I was waiting for an answer, and didn’t got any. Q. Then what? A. Then I sung out again. Afterwards I sung out to the man at the -wheel, ‘ Keep off,’ about the same time. Q. What happened after that, after you sung out to the man at the wheel to keep oíf? A. I don’t know what happened. I, kept looking at the vessel,—at the schooner. Q. What did you see of her? A. I didn’t see anything right away, only a green light, and some kind of dark spot; I couldn’t make it out. After she came a little further I could see it was a schooner.”

On cross-examination-

“Question. You kept your eyes fixed on the lights, did you? Answer. Tor a moment, and then I went across one or two times, and I looked up again, and couldn’t see it.”

(8) He subsequently states that at no time did he see both the lights of the schooner; and yet, if the schooner, by a change to the northward, *880shut in her red light, both must have been visible together for a time before the green alone was1 exhibited.

(4) The time when the green light is said to have been first recognized is important, and can be very approximately determined from the testimony as to the incidents occurring between that and the collision. Getting no answer to the first report of the green light, the lookout shortly reported it again, and almost immediately shouted “hard-up.” The second mate, coming just then out of the companion way, heard the order, and repeated it. The captain heard the second mate give that order. He jumped up immediately, and without stopping to look at the schooner, or to see her lights, went to the help of the wheelsman, and found the wheel “within a few spokes” of hard-up. The wheelsman says this was a very short time before the collision. When the order “hard-up” was given, the wheelsman saw the schooner’s red light under the foresail, “about two or three of her lengths distant,” i. e., from 200 to 300 feet; and, as the vessels were approaching each other at the rate of 1,100 feet per minute, this testimony xyould indicate that the green light was seen only from a quarter to a half minute before collision. The second mate, after repeating the order “hard-up,” hurried forward, and when he got to the forecastle, the brig’s jib-boom was just getting between the schooner’s masts. The first mate heard the second mate’s steps on his going to report to the captain, and he heard the order “hard-up ” very soon after; then he jumped out of his berth, came up on deck, and saw “the schooner’s jibs coming past the brig’s jib-boom;” he “looked astern and saw the captain at the wheel, and then made a leap forward as fast as he could go, and got a little more than amid-ships when they struck.” Upon all this testimony I cannot find that there was more than half a minute between the report of the green light and the collision. The respondent’s testimony leaves great doubt, also, whether the red light reported to the captain wras seen any considerable time before the collision, whatever that light was. When the lookout first reported it, the second mate was amid-ships, and came up on the forecastle. The lookout estimates that they talked about the light for about two minutes before the second mate went to report it to the captain. This is very improbable, if the light was readily distinguishable; probable enough, if-the light was so faint as to be doubtful. He subsequently gives two minutes, however, as his estimate of the time between seeing the green light and the collision. But as, upon all the testimony, there is no reason to believe that the latter interval exceeded half a minute, the former interval was quite as likely similarly exaggerated, and possibly not longer than the latter interval; and if the light was a clear red light, it is hardly probable that the second mate would remain with him even half a minute, instead of reporting it at once to the master, as ordered. The reasonable inference is that the red light was not seen more than a minute and a half, at most, before the collision, or about a quarter of a mile distant. This would convict the lookout of gross inattention, and his inattention would also be necessarily inferred from his not seeing both the schooner’s lights when she changed her course, if the red light was in *881fact the light of the schooner, and the green had been brought into view solely by such a change.

Another circumstance that diminishes confidence in the lookout’s testimony is his intimation that when the schooner's green light was first seen he could not see her hull or sails; he could only see “a green light and some kind of a dark spot. I couldn’t make it out. After she came a little further, I could see it was a schooner.” It was a bright, starlight night; it was already dawn; and, upon the other testimony in the case, it is very clear that at the time when, as now claimed, the green light was first visible, the vessels were so near, i. e., within 500 to 700-feet, that the hull and sails must have been in plain view. It is impossible, upon testimony of this character, to hold the libelant’s narrative a fabrication in its main particulars. See Chamberlain v. Ward, 21 How. 548, 562.

The Angle of Collision. Had both vessels collided without any change by either of their former courses, the angle of the blow would have been hut one and one-half points, if the schooner was in fact sailing N. by ID. I do not regard, however, the latter fact as certain, because she was not sailing by compass, but by the wind. The angle of collision, as shown in the drawings from the models placed by various witnesses to illustrate the collision, varies from four to six and one-half points. Hie master of the schooner places the vessels at an angle of four and three-fourths points, the mate at four and two-thirds, the steward at four points. But the master, who came on deck but a few seconds before the collision, states that the brig’s red light was then about two points, only on his starboard bow. All the other estimates of the bearing are the same. Moreover, as the brig was going nearly twice as fast as the schooner, and was from two to throe times the schooner’s tonnage, striking her forward of amid-ships, would necessarily, almost immediately send the schooner’s bows to the northward, and increase the apparent angle of collision. It is not at all improbable, in my judgment, and it is probably true, that within the last few seconds before the collison the wheelsman luffed somewhat in the hope of avoiding the brig, which was close upon them. The wind is stated as “about N. 1ST. ID.” I do not find that the precise course of the schooner, by compass, was noted. The fact that the sails of the brig, about 4 o’clock, were trimmed aft more than they had previously been, affords some indication that the wind was hauling to the northward; and, if so, the course of the schooner would gradually veer also to the northward. And it is noticeable that the steward, who could have had no idea of the bearing of his answer, gave his judgment of the schooner’s course, simply from general knowledge of the sound, as being N. E. or N. E. by E., though he did not pretend to swear to this; and, if the schooner was in fact sailing a point more to the northward than she had previously been, the hearing of the Prig’s light would he two points off her starboard how, as her witnesses estimate it ivas. Opposed to this is the mate’s statement that Faulkner Island light bore E. N. E., one point on his port how, about ton minutes before the collision. I cannot place much reliance upon the brig’s, *882testimony as respects the position of the schooner’s sails. The second mate says the foresail and mainsail were full on the starboard tack, as the sails were on the port side. The master says they were full, with the booms on the port side. Several say the jib-sheets were hauled to starboard. The lookout says the sails were fluttering “a little.” If the sails were full on the starboard tack, the schooner must have completed her tack, and changed her course from 9 to 10 points. Not only is it impossible that any such change could have been made in the short interval of a half minute from the time when the green light was first seen till the collision, but it seems to be absurd and incredible to suppose that the schooner, with the brig’s lights near at hand, on her weather bow, (as the respondent’s contention requires to be assumed,) should have tacked almost under the very bows of the brig, to cross the brig’s course, when she was safe in holding her own course to the leeward of the brig. Such a. change is denied by ail on board the schooner, and cannot be accepted as a fact. If, again, the schooner had luffed five or six points, the sails would have been shaking, not “a little,” as the lookout puts it, but violently. There is testimony on behalf of the brig that under her starboard wheel she fell off one and one-half points before the collision) so as to head W. by S. But from the short interval, I have much doubt whether she paid off a point. Although the respondent’s witnesses do not directly allege any change of course by the brig, and say they did not perceive any change in her course, the master and the mate judged that they must have been heading towards the north-west. But this is based upon the assumption that the schooner did not materially change her heading. Much less reliance is to be placed, I think, upon any estimate of the angle of collision, in the excitement of such an occasion, than upon the bearing of the lights a few seconds before, which seamen are accustomed to observe and to, estimate. Taking, however, all the testimony together, I judge that at the moment of collision the angle of the vessels was at least three to four points. But as the precise course of the schooner is not known, and as it is quite probable that the wheelsman put his wheel to starboard at the last moment, as he would naturally do when the brig ■was nearly upon them, I cannot give great weight to this element in the case. The testimony of the master that when he came on deck, probably not over 15 or 20 seconds before the collision, the brig’s red light bore about two points off his starboard bow is, I think, much more likely to be correct than his estimate of the angle of collision. What was observed afterwards when the schooner’s bows had naturally swung to the northward, might easily be taken by all the witnesses for the angle of collision. Most of the respondent’s witnesses did not see the schooner at the moment they struck; and hence cannot tell the angle of the blow. The'master of the schooner says that when he came on deck the collision was inevitable, and that he therefore gave no orders. If the sails were then shaking, he would have noticed it as he came out of the companion way. It is probable that the wheelsman, if the master gave no order, of bis own motion luffed when the brig was near, and after the master had come up. There was still time for her to change three or four points, *883as she was a small vessel; and the shaking of the sails on the close approach of the collision would not be likely to attract the attention of the men on board.

In the case of The Fairbanks, supra, it was sought to prove that the sailing vessel changed her course from the appearance and effect of the blow upon the steamer, which indicated a right angle. The court say:

“The circumstances adduced do not satisfy the court that any such change of course was made by the brig as is supposed, certainly not until the proximity of the two vessels was so close that a collision was inevitable; and then it is quite clear that the steamer made a sudden change, and it may be that the brig also changed her course, as is supposed by the claimants. Fault, under such circumstances, will not be imputed to the vessel required to keep her course, if she was otherwise blameless.”

In the present case, neither the angle of the blow, nor the luffing of the schooner to the northward, if such was the fact, is material, unless the schooner’s course had been previously such that her red light was exhibited to the brig; because, if the schooner’s red light was not shown to the brig, the latter was in no way misled; nor were any efforts on her part to avoid the schooner thwarted by the latter. Jf the testimony of the witnesses for the schooner is to be believed, and I see no reason why it should not bo, it is clear that the brig’s red light, seen off' the schooner’s starboard bow, did not draw nearer to the schooner’s stem, when the brig was approaching somewhat near. This is conclusive evidence that the brig was not moving on a path that would cross ahead of the schooner without collision; and in that situation the schooner’s starboarding, if the helmsman did starboard, was not material, and did not produce the collision, since it tended to take the schooner away from the line of the brig’s course. The question as to the angle of collision bears chiefly upon the credibility of the schooner’s witnesses, and, in my judgment, the evidence relating to. this angle is not sufficiently certain as to the amount of the angle, the time it was observed, and the causes of any such changes as may have taken place, to impeach the general credit to which I think the schooner’s witnesses are entitled.

My conclusions are: (1) That the schooner’s course was from E. by X. to E. X. E., and that her green light was at all times presented to the brig. (2) That the brig’s red light was at no time seen on the schooner’s port bow, or ahead, and consequently that the schooner’s red light was not shown to the brig at any time. (3) That any red light seen two and a half or three, points on the brig’s port bow was not the red light of the schooner. (4) That the schooner made no material change of course that misled the brig, or contributed to the collision; and that the green light reported by the brig’s lookout, not more than half a minute before collision, was not over a point or a point and a half on the brig’s port bow, and its being seen was not caused by any change in the schooner’s course. (5) That the collision was caused by the fault of the brig, in not seasonably observing the lights of the schooner, in not keeping a constant watch upon them, and in not taking timely measures to avoid her. (6) If the light first seen was the schooner’s light, *884it was not over a point on the brig’s port bow, and was the green light, and not the red light.

If it had appeared in this case, or been proved in any case, that the .green light, by reason of faintness after long burning, or the partial obscuration of the glass, or from the effect of the dawn, showed red at first, 1 should give that explanation to the respondent’s testimony in this case; and, in that event, should have held the schooner liable for such a defect in her light. Though a claim of such a different appearance in the green light has sometimes been made, I do not know of any case in which it has been established or recognized as a fact. Nor have I overlooked several conjectures which would admit of the brig’s actually seeing the red light of the schooner, and afterwards her green light; but every such conjecture is incompatible with the two facts on the part of the schooner which I consider established, namely, that the brig’s red light was seen all the time on the schooner’s starboard bow, and, as it approached, did not draw nearer her stem. Compelled to choose, therefore, between the narrative of the respondent and that of the libelant, I feel constrained to uphold the latter, and award a decree in favor of the libelant, with costs, and a reference to compute the damages.