This is an appeal from an interlocutory decree holding the steamship Yoshida Maru alone at fault for a collision with the McCormick, on the lower Columbia river, about 3 o’clock in the morning-of April 17, 1925. The Yoshida was proceeding downstream with a cargo of lumber, and the McCormick was coming up the river, partly loaded. The collision, which was almost end on, occurred at a point between Flavel and Port Stevens, where the channel, though comparatively narrow, was of ample width for safe passing. The night was dark, but not stormy, and generally it may be said that conditions for navigation were fair. The rules (article 25 [Comp. St. § 7899]) required each vessel to keep to the right side of the channel, if such course was not dangerous or impracticable. Unless relieved from so doing by necessity or the consent of the McCormick, it was therefore the duty of the Yoshida to proceed on the north side, but instead she sought and undertook to make a starboard passage without awaiting the McCormick’s consent.

The only justification attempted by her pilot in his somewhat inconsistent testimony is that he believed or suspected that in disregard of the rule the McCormick proposed to take the north side. Upon some of the material questions the testimony is highly conflicting, but we agree with the lower court that under the facts, either admitted or established by a clear preponderance, there was no warrant for his acting upon such an assumption. The truth, in all probability, is that, desiring the deeper water, he steered his course toward the south side, trusting that the McCormick would yield her right and avoid danger. Undoubtedly such is the import of some of the testimony, and that he believed the McCormick was of her own notion taking the north channel is highly improbable. The version of his testimony most favorable to the Yoshida is that, when he first sighted the McCormick, then below Port Stevens, she was showing her red light, but about the time she passed Port Stevens her green light came into view, and thereafter remained visible. This was at a distance of at least a mile and a half. He does not contend that she sounded her whistle or otherwise signified her intention to take the north side. The deeper, and hence the preferable, channel was on the south side, which of right belonged to her; and yet we are asked to believe that immediately upon observing her green light, a mile and a half away, he in good faith concluded that against interest and in violation of a fundamental rule she was going to take the north side, and that he feared, if he insisted on making a port to port passage, as was his right, he would be forced beyond the navigable channel. It was his duty to assume that the McCormick would observe the law of the road until there was at least substantial reason to fear she would not, or could not, do so. Any other practice would be prolific of disaster. The Albert Dumois, 177 U. S. 240, 249 , 20 S. Ct. 595 , 44 L. Ed. 751 . If, in truth, the McCormick did show her green light as she- cleared Port Stevens wharf, the fact would be of little significance. She was at that time more than a mile and a half away. The intervening fairway was clear, and, as has already been stated, the south was the better side.

The sounding by the Yoshida of the two-blast signal about the time she first observed the green light does not materially alter the ease. The testimony on the point by the McCormick’s officers is to the effect that it was not heard by them. But, however that may be, admittedly the latter did not respond, and without an answer the signal can be regarded only in the nature of a request to the McCormick, and established no right against her. The John King (C. C. A.) 49 F. 469 ; The Cygnus (C. C. A.) 142 F. 85 . “Where a vessel has no right to pursue a particular course without receiving the assent of' the vessel she is meeting, the whistles she uses to obtain that assent are merely invitations to an agreement contrary to the usual mode of passing. * * * ” The New York (C. C. A.) 86 F. 814 . We are convinced that in violating the plain rules of navigation the Yoshida was primarily at fault, and that without such fault the collision would not have occurred.

The remaining question, and the more •difficult, is whether the McCormick also was so far at fault as to justify an apportionment of the damages. It is well settled that, in case of a collision, the initial fault of one vessel does not exempt the other from the duty of complying with the rules of navigation, or of using such precautions as good judgment and good seamanship require to meet the emergency. The Maria Martin, 12 Wall. (79 U. S.) 31, 47, 20 L. Ed. 251 ; The Sunnyside, 91 U. S. 208, 213 , 23 L. Ed. 302 ; The New York, 175 U. S. 187 , 20 S. Ct. 67 , 44 L. Ed. 126 .

True, as we have already indicated, up to a certain point the preferred vessel has the right, and, indeed, it is her duty, to maintain her course and speed, and she is not to act upon a mere suspicion or surmise that the .approaching vessel will fail to do her duty. The Britannia, 153 U. S. 130 , 14 S. Ct. 795 , 38 L. Ed. 660 ; The Northfield, 154 U. S. 629 , 14 S. Ct. 1184 , 24 L. Ed. 680 ; The Delaware, 161 U. S. 459 , 16 S. Ct. 516 , 40 L. Ed. 771 . But, as was said in The New York, supra, “the fact that a steamer is entitled to hold her course does not excuse her from inattention to signals, from answering where an answer is required, or from adopting such precautions as may be necessary to prevent a collision, in case there be a distinct indication that the obligated steamer is about to fail in her duty.”

In view of our finding that the initial fault was with the Yoshida, we are also to be mindful of the principle that “where one vessel, clearly shown to have been guilty of fault adequate in itself to account for the collision, seeks to impugn the management of the other vessel, there is a presumption in favor of the latter, which can only be rebutted by clear proof of a contributing fault.” The City of New York, 147 U. S. 72, 85 , 13 S. Ct. 211, 216 ( 37 L. Ed. 84 ); The Oregon, 158 U. S. 186, 197 , 15 S. Ct. 804, 809 ( 39 L. Ed. 943 ); The Umbria, 166 U. S. 409 , 17 S. Ct. 610 , 41 L. Ed. 1053 ; The Victory, 168 U. S. 410, 423 , 18 S. Ct. 149 , 42 L. Ed. 519 .

By the court below it was considered that the evidence is sufficient only to raise a suspicion that the McCormick was guilty of contributing negligence, and respondent seeks to invoke the rule that in a doubtful case findings based upon conflicting evidence will not be disturbed on appeal. If applicable at all, the rule can have but little force, for the reason that, of the several officers and members of the crews of the two vessels giving their testimony, only one testified in the presence of the trial court, and that was the Yoshida’s pilot, whose version and explanations, we have already stated, are in some respects inconsistent and unsatisfactory.

But scarcely more satisfactory we are inclined to think is the testimony of Capt. Christensen, master of the McCormick, especially when it is considered together with his earlier statements, made in the course of hearings before the inspectors. There he admitted that, after the danger of collision had become imminent his management was faulty, and these admissions he did not successfully explain away in his later testimony. Such mistakes we may perhaps overlook, as having been made in extremis, and we turn to his earlier conduct. This we consider largely in the light of his own testimony and that of disinterested witnesses.

It is to be noted that the Yoshida did not abruptly change her course in passing from the north to the south side of the channel. According to Christenson’s version of what occurred, she must have been near the north side when she passed Flavel, about 5,000 feet easterly from the point of collision, and from that time she steadily pursued a diagonal or angling course toward the south side of the channel. Christensen observed no substantial change, for he testified that her green light showed continuously, and her range lights were open to the south. It is reasonable, therefore, to assume that she got south of the middle of the channel at least 2,000 or 2,500 feet east of the point of collision, a position suggestive of the possibility of danger. While Christensen testified that he did not hear the first two-blast signal from the Yoshida, or the second, if it preceded his one blast, in his experience signaling for a starboard passage in the lower reaches of the river was not uncommon, and he therefore knew that for one reason or another such a passage was sometimes made.

Observing the uniform course of the Yoshida toward the south side, it was properly a matter of the deepest concern that at the earliest possible moment he be advised of what she intended to do, either by appropriate signal or an actual change of course to her own starboard. So impressed was he with what he observed that, when the two vessels were a mile or more apart, he exclaimed in the presence of the second mate, who was on'the bridge with him, “My God! what is that fellow going to do ?” a clear indication of donbt upon his part as to the Yoshida’s intentions, a donbt thought to be sufficient to require caution and inquiry. By article 18, rule 3 (Comp. St. § 7892) it is required that “if, when steam vessels are approaching each other, either fails to understand the course or intention of the other, from any eause, the vessel so in doubt shall immediately signify the same by giving several short and rapid blasts, not less than four, of the steam whistle.” Such a signal admittedly the McCormick did not give, either immediately or at all.

Whether in itself this failure may be said to have materially contributed to the accident involves a consideration of testimony respecting, not only what occurred, but also the order of occurrence, as well as times and distances, which, even as given by the McCormick’s witnesses, including Capt. Christensen, is very indefinite and measurably confused and inconsistent. Broadly, appellees’ contention is that the doubt in Christensen’s mind did not arise until it was too late to do more than what was done to avoid the collision, and when the signal would not have apprised the Yoshida of any dangers of which she was not already aware. But Christensen’s exclamation clearly imports a very considerable uncertainty, and if, as he states,that was made when the two vessels were at least a mile apart, it is difficult to understand why, with the required signal and fair seamanship, the collision should then have been unavoidable.

Christensen testified that the exclamation was uttered after his one blast, and just before the ensuing two blasts from the Yoshida,' which he contends constituted the first and only signal of that type he heard from the Yoshida, and which, he said, was indistinct. The important matter of signals is one in respect to which the testimony is highly conflicting. The overwhelming weight of the evidence, we think, supports the view that, before any signal was 'given by the McCormick, the Yoshida had given two two-blast signals, with a considerable interval between them, thus twice signifying her intention or desire to take the south side, and that it was not until a short time after the second of these that the MeCormiek gave her first sig* nal, a single blast. And we are further of the opinion that these signals should have been heard by the McCormick, and, if heard, should, of course, have been promptly answered.

Upon the point the two groups of interested witnesses from the two vessels are directly opposed, one to the other, but the testimony of Capt. Wieklund, who is in a measure corroborated by pilots at Astoria, is convincing. He was the officer in charge of the Coast Guard station at Hammond, on the south side of the river, almost opposite the point of collision, which position he had occupied for 28 years. That he was both competent and disinterested admits of no doubt. Owing to the illness of his wife, he was awake, and he was interested in listening to signals on the river, for the reason that he was expecting to be saluted by the engineer of an outgoing motorboat, who had at one time been a member of his crew. He testified that he first heard a two-blast whistle, up near Flavel he imagined, and then two blasts from the same vessel when she was further down toward his station, then a long blast from a vessel further west, quite a distance from the first vessel, and then three- short blasts of the whistle of the first vessel, and lastly several short blasts from the second vessel. He was in the house at first, and took no action until after the two two-blast signals, but, upon hearing the long blast from the second vessel, he concluded a collision was imminent and jumped out of bed, put on his clothes, and called out his crew. The three short blasts of the first vessel he heard while he was dressing, and these he understood to be a signal for full speed astern. While perhaps they are not so clearly disinterested or dependable,, pilots waiting to, go on duty at Astoria, three or four miles up the river, and against the wind, substantially corroborated Capt. Wieklund.

In view of this testimony, supporting that of the witnesses of the Yoshida, it is difficult to believe that the MeCormiek failed to hear either of the two-blast signals from the Yoshida, or, if they did not hear, to understand why, if her officers and crew were competent, properly stationed, and attentive, they so failed. The only explanation suggested in the record is to be found in the fact that there was an abnormal amount of noise upon the MeCormiek and there was no lookout forward. -The McCormick’s second mate, Nelson, testified that he could not state positively that the Yoshida failed to give signals other than those he heard, “because our safety was blowing all the time.” Rocha, the quartermaster, testified that they were prevented from hearing other whistles by “the noise of the machine.” Encina, said to be the lookout, testified, that before the collision the McCormick’s boilers blew off six times. The only explanation given by Christensen of why he could not hear what Wicklnnd heard, though the latter was in the house, is that a strong breeze was blowing. These several witnesses were all stationed upon the bridge. The lookout had been taken off the forecastle head when the McCormick came into the river, and at the time of the accident there was no one forward of the bridge, which was approximately 125 feet from the stem.

By article 29 of the Inland Rules (Comp. St. § 7903) it is provided that, “nothing in these rules shall exonerate any vessel * * * from the consequences of any neglect * * * to keep a proper lookout. * * * ” And in the absence of special conditions, which did not here exist, the place for a lookout is at the bow. British Col. Mills Tug & Barge Co. v. Mylrole, 259 U. S. 1 , 42 S. Ct. 430 , 66 L. Ed. 807 ; Edward St. John v. Paine, 10 How. 557 , 13 L. Ed. 537 ; Chamberlain v. Ward, 21 How. 548 , 16 L. Ed. 211 ; The Ariadne, 13 Wall. 475 , 20 L. Ed. 542 ; The Ottawa, 3 Wall. (70 U. S.) 268, 18 L. Ed. 165 ; The Buenos Aires (C. C. A.) 5 F.(2d) 425 ; Wilders S. S. Co. v. Low (C. C. A.) 112 F. 161 . In respect to seeing, the improper position of the lookout was without consequence, for from the bridge the men probably could see all that could have been seen from the bow. But clearly, under the testimony, the conditions forward 100 feet from the bridge were more favorable for hearing. Capt. Christensen testified that from his position on the bridge he heard the two-blast signal of the Yoshida but faintly, and that then the two vessels were so close together that collision was probable, if not absolutely unavoidable. How, then, can we hold the “McCormick’s” management to have been reasonably skillful and prudent, when she could not, until too late to escape imminent peril, hear the Yoshida’s signals, which'were distinctly heard in the house by Capt. Wieklund, and were at least audible three or four miles up the river ?

Reprehensible though we may think the conduct of the Yoshida to have been, in attempting a starboard passage before getting the express consent of the McCormick, upon a careful consideration of the whole record, we are constrained to the conclusion that the latter is also chargeable with negligence contributing to the accident, and accordingly the decree below is reversed, with instructions to enter a decree in harmony herewith, and for further proceedings; cost of appeal equally divided.