TAFT, Circuit Judge
(after stating the facts). We must first decide what are the rules of navigation to which, the colliding vessels were obliged to conform. The collision, occurred in Canadian waters, and it is contended by counsel for the appellee that the Canadian statute of navigation must govern the court in the consideration of the conduct of the parties. It is settled by the decisions of this court in The North Star, 22 U. S. App. 242, 10 C. C. A. 262, and 62 Fed. 71, and The City of Mackinac, 43 U. S. App. 190, 20 C. C. A. 86, and 73 Fed. 883, that, in the absence of proof of the Canadian statute, the proper navigation at the time of this collision was prescribed by section 4233 of the Revised Statutes of the United States, as supplemented by the rules adopted by the supervising inspectors under the authority of section 4412, Rev. St. It is conceded that *822at the hearing in the court below the Canadian statute was not introduced in proof, and that neither the counsel nor the court relied on its provisions. It is also apparent from the evidence that the captains of the colliding vessels both regarded themselves as acting under and subject to the federal statute and the supervisors’ rules at and before the time of the collision, xlt the hearing of the motion made by libelant for a rehearing and a modification of the decree so as to hold the Conemaugh free from fault, some reference seems to have been made to the Canadian statute. This we gather, not from the record, but from, the affidavits of the counsel for libel-ant, and the clerk of the district court, filed in support of a motion for a certiorari. From the affidavit of the clerk it is to be inferred that the reference to the Canadian statute was only arguendo, and that there was no formal offering of the same in evidence. Indeed, it is difficult to understand how there could have been an offering of the same as evidence upon the issue made on the pleadings, because the action of the court in modifying the interlocutory decree seems to have taken on the evidence as adduced at the trial, and without a new hearing of the cause. The motion of libelant for rehearing asked for leave to introduce new evidence, but the Canadian statute was not mentioned in the description of the evidence to be offered. The respondent asked leave to introduce new evidence after the court had modified the decree, and this was denied. Now, the respondent had stood upon the evidence of libelant at the trial, and had adduced no evidence of its own. If the libelant had been permitted, on a rehearing, to introduce the Canadian statute, ándito change materially the rules of conduct to which the parties were to be held, then it would seem hardly fair not to have allowed the respondent to call its witnesses to meet a different case from that in which it had not deemed it necessary to call any one. But, disregarding these considerations, the conclusive reason why the court cannot consider the Canadian statute as part of this record is found in the return of the district court to the writ of certiorari. It contains no certificate that the Canadian statute was made part of the record by being offered and received in evidence, but only a statement by the clerk that that which is returned is a correct copy of the Canadian statute, as published. The district court and the clerk seem to have construed the action of this court in issuing the writ as a decision or finding that the Canadian statute was a part of the record below, and an order to certify the same, whereas the writ merely directed the court to complete the record if, in any respect, it was defective, leaving to that court to decide what constituted its record. We cannot regard the Canadian statute, therefore, as in evidence, or as part of the record before us. It might have been a question, even if the Canadian statute had been properly proved, whether two merchant vessels of the United States, proceeding from one port of the United States to another, and incidentally crossing and recrossing the national boundary, were not, though in Canadian waters, still to be held by a court of the United States as bound by section 4233, the opening words of which are as follows: “The following rules for prevent*823ing collisions on the water shall be followed in tlie navigation of vessels of the navy and of the mercantile marine of the United States.” We do not decade this point, because; though suggested by counsel, it is not before us. All that wc do Isold Is that, in the absence of the proper proof of the Canadian statute, the presumption is that section 4233 and the supervising inspectors’ rules furnish the law of navigation for the cause.
It is not disputed that the courses of the two vessels were crossing. so as to involve risk of collision, and that the Conemangh had the New York on her own starboard side. Under such circumstances, by rule 1!) of section 4233, Kev. St., the Conemaugh was required to keep out of the way of the New York; and by rule 23, the New York was required to keep her course, unless, as provided in rule 24, special circumstances existed, rendering a departure from rule 22 necessary to avoid immediate danger. Buie 2 of the supervising inspectors further limited the discretion which the Conemaugh had in selecting the manner in which she could keep out of the way by providing that when steamers were approaching each other in an oblique direction, as these were, they should pass to the light of each other, as if meeting “head and head,” or nearly so. The learned district, judge was of opinion that, rule No. 2 did not apply in this case, because he thought the situation here was within an exception to rule 2 stated in a note to the supervising inspectors’.rules, bv which all the rules are made inapplicable to steamers navigating in a crowded channel. In this we cannot agree with him. The width of the navigable channel between the tow and the Canadian shore before and at the time of the collision was variously estimated as from 500 to 750 feet. For reasons which we shall hereafter stale, we think it was about 500 feet. The Conemaugh had not entered that: channel, but was above it in the river at least 300 feet. She had the whole width of the river on her st.arboa.rd hand, and had full opportunity to port her helm and run down into the bight of the tow, out of any danger, bad she desired to do so, and this with very little delay. Had she done this, there would have been no collision. It follows that she was guilty of a fault which caused the collision. We should have reached Hi is conclusion, even if the Conemaugh was not bound by rule 2 of the supervising inspectors, and was only under obligation to keep out of the way of the New York, with discretion to pass her on either hand. The evidence satisfies us that the Conemaugh was in the course of the New York when the collision occurred. What was the course of the New York? Her general course was upstream, and probably, if she followed the usual track of steamers (though this was not invariable), a little towards the American side of midchannel. It is well settled, however, that a vessel does not depart from her course when she turns from her general course to avoid obstructions, of which the vessel keeping out of her way must know the existence and must: allow for the effect. The Iron Chief, 22 U. S. App. 473, 11 C. C. A. 196, and 63 Fed. 289; The John L. Hasbrouck, 93 U. S. 405; The D. S. Stetson, 4 Ben. 508, 7 Fed. Cas. 1132; The John Taylor, 6 Ben. 227, 13 Fed. Cas. 896; The Velocity, L. R. 3 P. C. 44; Mars. Mar. Coll. (2d Ed.) 473.
*824The proper course of the New York was that which the Conemaugh ought to have known she would naturally have taken had the Cone-maugh not been in sight. As the New York came up the river the Burlington’s tow was stretched across the river, and by an exchange of single blasts a proper agreement had been reached, by which the New York was obliged to go round the tail of the tow, having it on her port hand. This required the New York, coming up on the American side of the channel, to port her wheel and change her course towards the Canadian shore. As she was about a mile distant when the signals were exchanged, it is highly probable that she could not, in a dark night, at once determine the length of the tow, or fix the place of the last barge in it. It was entirely natural and proper navigation for her to change her course only moderately to starboard until she could pick up the tail of the totv, and avoid going uselessly near the Canadian shore, and more out of her general course up the river than necessary. The evidence shows, then, that after first porting her wheel she ran on a course which would have carried her into the Amaranth, the third barge in the tow; that when about 800 feet away she ported again, and took a course which was about parallel with the then course of both the Amaranth and the Ferguson, the last two barges of the tow, and 100 feet distant therefrom, towards the Canadian shore. Their course was about two or three points towards the American shore from the course of the river and channel,, and so the course of the New York was then two or three points from the midchannel line towards the Canadian shore. The great weight of the evidence establishes that the NewT York did not again change her course to starboard after she ported her wheel 800 feet away from the tow to pass the last two barges. The libel charges that when near the last barge she ported her wheel and swung violently to starboard, and thus brought about the collision. We think the evidence utterly fails to show this, and that she made no change of her course to starboard which the presence of the tow, sagging downstream, and slowly crawling across the river, did not make necessary. But it is said that after the New York passed the tow her proper course was to swing to port under the stern of the last vessel in the tow, and thence over towards midchan-nel, instead of which she continued on towards the Canadian shore, and ran into the Conemaugh. It is undoubtedly true that the New York’s proper course, after passing the tow, was to resume her general course upstream near midchannel. The John L. Hasbrouck, 93 U. S. 405. All the witnesses who observed her course admit that just before the collision she was swinging under a starboard wheel. It would seem, therefore, that she had begun to change her course to port; and the only question is, did she begin to do this as soon as she ought to have done it? How soon ought she to-have done it? She was not obliged to turn a sharp corner round the stern of the last barge on the tow. She certainly would not have done this had the Conemaugh not been there, and, as we have seen, her proper course could not be affected by the fact of the Conemaugh’s presence. Her natural course would have been to swing gradually to port, under a slowly-turning starboard wheel, so as to make an easy sweep back to *825midchannel. The Oonemaugh could not, by pressing on it, make the course of th.e New York one requiring her to dodge in between the tail of the tow and the Oonemaugh. In answering the question whether the actual course of the New York was in accord with her proper course thus stated, we may derive considerable light from the evidence as to the distance of the New York and the Oonemaugh from (he tow when the collision occurred. The captain of the Oonemaugh says (he distance was 750 feet. The captain of the Ferguson puts it at about the same distance. The mate of the Oonemaugh. makes the distance about 000 feet, and this is the effect of the evidence of the captain of the Amaranth. We are of opinion, from the circumstances in the case, that the smaller distance is much more likely to be correct. An examination of the chart shows that the distance which the Oonemaugh was from the tail of the tow, when her engines were checked, did not exceed 1,200 feet. The statement of her engineer as to the time between that check and the collision was about four minutes. The relative speeds of the Oonemaugh and the tow were such that the former was gaining on the latter at least three miles an hour. Allowing for the curved course the Oonemaugh took in following the tow down, and allowing half a minute between the hard a-starboard swing of the Oonemaugh and the collision, she must certainly have been within .200 feet of the tow when the swing occurred. But it is said that this conclusion is at variance with the place where the Oonemaugh grounded on the channel hank. The weight of the evidence shows that the distance of the tow from the Canadian shore was about 750 feet. The channel bank was about 235 feet from shore. This left the channel between the tow and the hank about 515 feet. It is not clear just how much time there was between the starboarding of the Coneinaugb's wheel and (he collision, but if. is quite evident ihat. there was some time in which to make headway towards the Canadian shore, and after the collision the evidence is that the engines of the Oonemaugh worked1 ahead strong for a minute or more. This is quite sufficient to show that the Oonemaugh might have starboarded her wheel at a point about as far from the shore as was the tow, and though when she blew her alarm signal she was but 300 feet from the tow, that after being struck by the New York, and working hard towards the Canadian shore, she might have brought up on the bank at a point much further than 300 feet from the position of the tow, which had been constantly moving away from the point of collision. Wifh the distance between the Oone-maugh and the tow but 300 feet., where was the course of the New York with respect to them? It is clear to us that the course of the New' York vrould not naturally be confined to swinging on her starboard wheel through the passage not much wider than her length. That wrould not have been the easy sweep which she was entitled to make in turning back towards midchannel. The Oonemaugh, therefore, being where she was, was either in, or dangerously near, the course of the New York, and was not keeping out of her way. More than this, she increased her fault by throwing herself right across the bows of the New York. The point where the New York struck *826her, to wit, only 30 feet from her stem, shows that if, when she blew her alarm whistle, she had ported her helm, instead of starboarding, she would have entirely avoided the New York, by passing that vessel port to port. It is very difficult to explain the navigation of the Oonemaugh, or to reconcile some of the statements of the captain of the Oonemaugh with the admitted situation of the vessels. He says that when he steadied, after swinging round the Kasota spiles and heading across the river, he saw the red light of the New York coming up the river, and whistled two blasts to her; that as he swung slowly around on his port wheel, following the tail of the tow, he saw both side lights of the New York; and that he continued to do so when he whistled his second signal of two blasts and his third signal of two blasts. An examination of the chart and the necessary courses of the two vessels makes it impossible that he could have seen the green light of the New York from the second to the third blast. The New York was proceeding from the American side in a slanting direction across the river, while the Oonemaugh was proceeding down the river in a slanting direction, and each must have been showing to the other but one light. We did not hear in the argument of counsel, nor can we find in the briefs, any explanation of how both lights of the New York were so long visible to the Oone-maugh as her captain testifies. The other witnesses from the Oone-maugh differ with the captain as to the lights shown by the New York. Hogan, the mate, saw only her red light at the time of the third signal, and Crowe saw her red light immediately after the second signal. The only importance of this error in the testimony of the Gonemaugh’s captain is that it shows that the porting of the New York’s wheel twice in her course from the American side to the tail of the tow must have been evident to him before there was any danger of collision. Another circumstance of much significance in this case is the course of the Oonemaugh with respect to that of the two last barges of the tow. Both barge cajffiains say that for some time before the third signal blast they saw both side lights of the Oonemaugh. Now, this is only possible if the Oonemaugh, instead of crossing their wake, was following along in it for an appreciable time while she was blowing the last one, and probably the last two, of the signals to the New York. It thus appears that, while she was blowing signals indicating a purpose to pass the New York starboard to starboard, she was continuing on a course port to port of the New York. It is not explained why, if the captain of the Oone-maugh regarded himself as having the right to select his course', and intended, as he says he did, to pass on the Canadian side of the New York, he did not direct his vessel towards the Canadian shore at once;, instead of following the tow down on the American side of the course which the New York must follow to dear the barges, and then suddenly swinging across the bows of the New York when that vessel Avas so near that collision was inevitable. On the whole case, Ave are clear in the conclusion that there wrere several glaring faults in the management of the Oonemaugh, AAhich caused the collision.
The question remains, was the New York also guilty of faults in *827navigation contributing to Ibis collision? Her owner, in its answer, admitted that those in.charge of her neither saw nor heard the Oone-maugh until the third double-blast signal, and iliat then they only heard the signal, without seeing the vessel giving it, and so supposed that the signal was not intended for them. The witnesses for the Conemaugh unite in saying that the New York was going at a speed of 10 mile's an hour, and apparently not under check. The district court found, from her failure to see and hear the Conemaugh, that the New York’s lookout was defective, and that she thus committed the fault of not; answering the Conemaugh’s signals. He further held that she should have checked or stopped when there was risk of collision, and that her failure to do so was a fault contributing to the disaster. It must be conceded that, if the New' York bad heard the signals of the Conemaugh, she would not have been obliged to respond by a double blast, and signify her willingness to pass starboard i.o starboard, instead of port to port. Under rule 2 of tie supervising inspectors, tlie Conemaugh was obliged to keep on tie New York’s port hand, and nothing but her consent — expressed in a double blast —to depart from the rule would justify the Conemaugh in assuming that consent. This was not a case where silence gave consent. Rule 2 requires signals to be given and promptly returned. It has been suggested that this requirement of a prompt answer applies only to the case‘where the first signal is to indicate a compliance with the rule, and not, as in the case at bar, where the signal invited a departure from it. This suggestion finds some support, it. is said, in the language of Mr. Justice Brown in The Delaware, 161 U. S. 459, 16 Sup. Ct. 516, where, speaking for the supreme court, he limits the use of signal blasts by the preferred vessel of two crossing vessels to an announcement that; she is maintaining her course according to rule. The learned justice said:
“These rules, however, so far as ttiey require the whistle to be used, are applicable rattier to vessels meeting end on, or nearly end on; and the. signals therein provided for are designed to apprise the approaching- vessel of the intention of the steamer giving- the signal to port or starboard, as the case may be. *\s applied to vessels upon crossing- courses, however, it means, when a. single blast is given by the preferred steamer, nothing more than that she intends to comply with her legal obligation to keep her course, and throw upon the other steamer the duty of avoiding her.”
We do not find it necessary to decide whether the New York should have returned a signal of one blast or not, because it is clear to us that her failure to do so did not contribute to tlie collision. So far as the Conemaugh was concerned, the New York’s silence was exactly equivalent to her express refusal to consent to depart from the rule by a single blast. There are several cases in which the point has been decided. The John King, 1 U. S. App. 64, 1 C. C. A. 319, and 49 Fed. 460, was a case of crossing vessels, in which the preferred vessel was condemned by the district court for not promptly returning an answer to a signal inviting her to depart from rule 2, The circuit court of appeals of the second circuit reversed this decree, and Judge Wallace, in delivering the opinion of the court, said, -referring to the other vessel:
*828“It was her duty, under sailing rule 19, to keep out of the way, and the duty of the ferryboat to keep her course. The red light of the ferryboat was plainly visible to the propeller, and there was nothing in the way to prevent the latter from passing astern of the ferryboat. She had concluded previously to pass across thé bow of the ferryboat, but had received no consent from the ferryboat to such a course, and there was still time to abandon that purpose and go astern. The latter course was plainly safe; the former, doubtful; and, quite irrespective of any rule of the supervising inspectors, common prudence required her to adopt the safe course and pass astern. She cannot invoke 'the aid of any rule of the supervising inspectors to justify her departure from duty without showing that her proposition to depart was heard, understood, and accepted by the ferryboat. If, by her signals, she invited a departure from the ordinary rules of navigation, she took the risk, both of her own whistles being heard, and, in turn, of hearing the response, if response' was made.”
Again, Judge Wallace says, speaking of the ferryboat:
“The signal she gave to the propeller when she got out into the river was the proper signal, viz. one blast, -to indicate that she proposed to keep to the right. If she had heard the second signal of the propeller, she could have done no more by way of a proper answer, and would have been under no obligation to give a different signal. This signal was given at a time when there was yet opportunity for the propeller to alter her course to starboard and pass astern. If we should assume that she heard the propeller’s signal, or ought to have heard it, and should have answered it by two blasts of her whistle, we do not see how the propeller was misled by the conduct of the ferryboat. We do not think, however, that, if the ferryboat had heard the propeller’s signals, her failure to answer them would have been culpable. The case, in its legal aspects, is quite similar to that of The B. B. Saunders, 23 Blatchf. 383, 25 Fed. 727, in which the court used this language: ‘Notwithstanding the inspectors’ regulations, therefore, the pilot of the Saunders was not bound to assent to the movement proposed by the Orient, unless due regard to the particular circumstances of the situation required a departure from the ordinary rule. Consequently, his failure to answer the signal of two blasts of the whistle from the Orient was not culpable, unless it was apparent that tire Orient could not safely pass astern of the Saunders.’ ”
See, also, The Florence, 68 Fed. 940; The St. John, 7 Blatchf. 220, Fed. Cas. No. 12,224; The Milwaukee, Brown, Adm. 313, Fed. Cas. No. 9,626.
. It is manifest to us that the failure of the New York to respond by a one-blast signal to the two blasts of the Conemaugh had no causal relation to the collision, because the silence of the New York was full notice to the Conemaugh that she must obey rule 2.
Again, how did the New York’s failure to see the Conemaugh contribute to the collision? Suppose the New York’s lookout had seen every maneuver of the Conemaugh; would her course have been different from what it was? We do not think so. She had the right and duty to maintain her course, and that we have found that she did. She would have had no right to infer that the Conemaugh would suddenly cross her bows, however alert her watch. She would have been justified in supposing that the Conemaugh, not having established an agreement to pass starboard to starboard, would maintain her bearing to the port of the New York, and swing clear on that side. Especially is this the case when, if she had seen the ■Conemaugh, she would have observed her swinging slowly to the port' of the New York, in the wake of the barges in the tow, although *829blowing signals of her intention, if assented to, to change her course io the starboard of the New York. But jt is said she ought to have stopped and reversed when there was risk of collision. The only risk of collision would have been in the Conemaugh’s failure to keep to the port hand of the New York, and this failure she was not bound to anticipate. The law on this subject has been settled by the supreme court in The Britannia, 153 U. S. 130, 14 Sup. Ct. 795, and The Delaware, 161 U. S. 459, 16 Sup. Ct. 516. In the latter case Mr. Justice Brown, speaking for the supreme court, used this language:
“The duty of a steamer having' the right of way, when approaching another steamer charged with the obligation of avoiding her, has been the subject of much discussion both in the English and American courts. That her primary duty is to keep her course is beyond all controversy. It is expressly required by the nineteenth rule of the original International Code (Rev. St. § 4233), and of the sixteenth rule of the Revised Code of 1885, and doubtless applies so long as there is nothing to indicate that the approaching steamer will nor, discharge her own obligation to keep out of the way. The divergence between the authorities begins at the point where the master of the preferred steamer suspects that the obligated steamer is about to fail in her duty to avoid her. The weight of English, and perhaps of Americau, authorities, is to the effect that, if the master of the preferred steamer has any reason to believe that the other will not take measures to keep out of her way, he may treat this as a ‘special circumstance,’ under rule 24, ‘rendering a departure’ from the rules necessary to avoid immediate danger.’ Some even g'o so far as to hold it the duty of the preferred vessel to stop and reverse when a continuance upon her course involves an apparent danger of collision. Upon the other hand, other authorities hold Thai: the master of the preferred steamer ought not to be embarrassed by clonbis as to his duty, and, unless the two vessels be in extremis, he is bound to hold to his course and speed. The cases of The Britannia, 153 U. S. 130, 14 Sup. Ct. 795, and The Northfield, 154 U. S. 629, 14 Sup. Ct. 1184, must be regarded, however, as settling the law that the preferred steamer will not be held in fault for maintaining her course and speed, so long as it is possible for the other to avoid her by porting, at least in the absence of some distinct indication that she is about 1o fail in her duty. If the master of ilie preferred steamer were at liberty to speculate upon the possibility, or even the probability, of the approaching steamer failing to do her duty aud keep out of his way, the certainty that the former will hold his course, upon which the latter has a right to rely, and which it is the very object of the rule to insure, would give place to doubts on the part of the master of the obligated steamer as to whether he woTÜd do so or not, and produce a timidity and feebleness of action on the i>art of both which would bring about more collisions than it would prevent. Belden v. Chase, 150 U. S. 674, 14 Sup. Ct. 264; The Highgate, 62 Law T. (N. S.) 841, 6 Asp. 512.”
This clearly shows that the New York had the right to maintain her speed, as well as her course, unless there was to, her some distinct indication that the Conemaugh was not going to keep out of her way by porting. She received no such distinct indication until the Cone-inaugh suddenly starboarded her helm and swung across the fast-approaching bows of the New York, and then it was too late to avoid the catastrophe. We find, then, that it was the fault of the Conemaugh which alone caused this collision, that the libel of her owner shohld therefore be dismissed, and that on the cross libel of the owner oí the New York a decree in personam against the owner of the Conemaugh for the agreed damage to the New* York should be entered. This conclusion disposes also of the petition of the inter-*830veiling insurance companies, which must also be dismissed. The decree of the district court is therefore reversed, with directions to enter a decree in accordance with these conclusions.