(after stating the facts as above). [1] Our study of the record has convinced us that both steamships *583were in fault. The district judge had the advantage, it is true, of seeing and hearing the witnesses. The investigation before him took a wide range both as to matters of fact and expert opinions. The steamships, the portion of the Detroit river and its sailing lines and lights which were involved after the ships each became concerned with the navigation of the other, the ads and omissions of the controlling officers and men of each ship during that period, and the relative steering qualities of the vessels, together with expert opinions as to means and methods of controlling them (especially the Fisher) under conditions stated and claimed by the respective crews, were described in great detail; and the cargoes were incidentally shown both in kind and tonnage. The testimony of the two navigating crews resulted as usual in two distinct and opposed theories; and, of course, if each ship had been navigated along the course her crew described, the admitted accident could not have taken place. In these circumstances the learned trial judge deemed it safer to determine which theory was “the more reasonable or most likely to be true,” than to pass upon the credibility of the witnesses. It must be conceded that the problem was not one of easy solution. Some of the facts, however, are not in material controversy, and we think they inevitably lead to controlling consequences.

1. Sailing Conditions Favorable. The navigating conditions were entirely consistent with a safe passing of these ships. Admittedly, the night was clear, and the lights of the ships and on the adjacent shores were alike easily discerned; there were no intervening boats to disconcert the eye and no wind or noises to disturb the hearing; the channel with abundant water was 800 feet in width; and the sailing line was straight and in the center of the channel between the points at which the boats were respectively sighted. In spite of these conditions, surprising difficulties arose soon after each boat had picked up the other. The Fisher was bound up the river and the Clement bound down.

2. Passing Agreement. The first difficulty of the case is met here. According to the master- of the Fisher, immediately after making the turn from the south to the north Grosse Isle channel range, he sighted a vessel ahead and a little below the Mamajuda lights, which proved to be the Clement, and received from her a one-blast passing signal, to which he responded with a like signal; and an agreement for passing port to port was thus concluded. The master of the Clement admits that a port to port passing agreement was made, though not through an initial signal of one blast from the Clement. He testifies that something else happened before. In one instance, he said when the Clement was above the Mamajuda lights, and, in another instance, when she was a little westward of the north channel Grosse Isle range, he “noticed the Fisher coming up the bend at Fighting Island,” and that when she crossed the intersection of the south and north Grosse Isle channel ranges he gave the Fisher “two whistles,” but got no answer, and after waiting “a sufficient time” he “blew two more.” The master of the Fisher declares he did not hear these signals; but however this may be, the master of the Clement thereupon noticed from her lights that the Fisher was on his starboard, and again that she had turned to her *584starboard so as to open her red light to the Clement, which, of course, means that the Fisher was then nearly parallel to, if not on, the north range. He says the Fisher then blew him “one whistle,” but that the Clement had not previously given a one-blast signal. Upon this subject the master of the Clement made several statements. In his direct testimony he said:

“When he blowed me the one I checked my steamboat down at once to slow speed and blew him an alarm or an attention whistle, whdtever you might call it, to attract his attention, and then as he had gone further over to the eastward, and the boats were so there was no danger, plenty of room to pass by, I answered with the one whistle.”

On cross-examination he stated that he thought it safe to accept the Fisher’s one-blast signal at the time it was given, but also said:

“I blowed a danger signal to attract his attention, and then answered it to let him know I understood he wanted that side, and was going to let him have it. Q. In other words, you said to him, ‘That is dangerous, old fellow, but you can have it if you want it’? A. Yes, sir. Q. What did you blow a danger signal then for? A. To attract his.attention. If he had understood me, I said, ‘Now, if you want that side, take it, all right, but understand it is dangerous.’ Q. To look out for yourself? A. Yes, understand w.e were agreed on it.”

It will be observed that the masters placed the two ships at substantially the same points in the Detroit river at the time each sighted the ship of the other; that they are in harmony as to the exchange of a one-blast signal; but that they differ as to the initial signal, each saying that his one-blast was given in response to a like signal from the other. Further, while both state that when the one-blast signals were exchanged the ships were in situations safely to pass port to port, yet it is plain that the ships could not then- have been as far apart as the masters say they were when sighted. At that time they placed the Clement near the Mamajuda lights and the Fisher at the crossing of the two Grosse Isle channel ranges. That portion of the north channel range is 6,000 feet in length. The master of the Fisher testifies that, when the ships exchanged one-blast sign’als, the Clement was in the neighborhood of three-quarters of a mile away, while the master of the Clement testifies that when the Fisher sounded one blast the Clement was within “five or six lengths” of the Fisher — “2,500 or 2,600 feet or more.” Such estimates of distance, it is true, are not very reliable. Yet these masters were men of long experience in navigating the Great Lakes and the Detroit river, and it is hard to conceive that they could have so mistakenThe length of the portion of the range they were sailing. This is especially true of the master of the Clement, since he .estimated the distance in lengths of his ship and in close correspondence with the distance he gave in. feet. We are therefore constrained to believe that the Clement, if not also the Fisher, was well advanced on the Grosse Isle north channel range when the exchange of one-blast signals' was concluded; and that the masters are so much at odds concerning both the distance between their ships and the circumstances which led to the exchange of single blasts as practically to destroy the value of their opinions touching the safety of passing port to port. Above all, the repeated, admissions of the master of the Clement that *585such a passing would be dangerous ought not to escape consideration when it is sought to ascertain the real causes of the collision. Such admissions reveal the effect made upon the mind of an experienced navigator at the inception of the conditions which so quickly culminated in disaster. This is not to exculpate the Fisher. It is in part, to test the soundness of the claim of the Clement that she was in no respect at fault. It is not an uncommon thing, it is the rule, to look for causes having their origin at some appreciable and material time before the collision, rather than to search for causes occurring after disaster becomes inevitable. Mr. Justice Clifford reiterated the rule in The Sunnyside, 91 U. S. 208, 209, 23 L. Ed. 302:

“Inability to avoid a collision usually exists at the time the collision occurs; but it is seldom a matter of much difficulty to trace the cause of the disaster to some antecedent omission of duty on the part of one or the other, or both, of the colliding' vessels.”

3. The Collision. Although the master of the Clement did not explain why he believed the port to port passing araugement would be dangerous, it is to be presumed that the conditions were such as to require most careful maneuvering of the ships in order to avoid disaster. In seeking causal connection then between the passing arrangement and the collision, it will be helpful to consider what the relations of the ships were at that time to the channel range, what was then done by both sets of navigators to avert danger, and also the speed of the ships. The navigating officers of the P'isher claim that when the passing arrangement was concluded they had the Grosse Isle north channel range lights astern of the Eisher, and the ship, which turned out to be the Clement, nearly dead ahead, whereupon an order was given and executed to port the Fisher’s wheel, and shortly after this order was repeated and obeyed for the purpose of opening the Clement’s range lights. The master of the Clement (in connection with a statement that he had shortly before passed certain other ships port to port) said the Clement “was on the north channel Grosse Isle ranges, or a little to the westward, or had been on the range”; again he and his second mate said the compasses of the Clement were then compared and adjusted and a reading taken with relation to the chart course, the one saying this was before the Clement reached the intersection of the Grosse Isle north channel and the Grassy Island ranges, and the other that it was done on the Grosse Isle north channel range; and it may be added that this location of the Clement agrees with the testimony of the navigating officers of the Fisher that when they got the channel range lights over her stern they had the Clement nearly dead ahead. Further, the master of the Clement says that, when the one-blast passing signals were concluded, the P'isher was over his port bow from 150 to 200 feet. Pie states, however, that he gave no order to his helm at that time, because he did not consider that there was any occasion for such an order. Plainly, then, whether under the testimony the Clement be treated as sailing on the range line, or on a parallel line to the westward, at the time it is said she accepted the Fisher’s one-blast signal, the distance between either line continued and the con tinuation of that of the Fisher, and consequently the degree, of safe» *586ty or danger in sailing the ships to the point of passing does not definitely appear. The most that can be said of the statement of the master of the Clement that the Fisher stood over his port bow, as stated, is that it tends to corroborate the claim of the Fisher that she ported twice after the' conclusion of the passing arrangement; but the reason stated by the master of the Clement for failure to give any order to her helm at that time is not convincing. We have seen that he claims to have checked the Clement down to slow speed and blown an alarm before he accepted the Fisher’s signal for a port to port passing. He does not claim, however, to have reduced her speed “to bare steerageway” (Navigation Rule 26, 28 Stat. 649); nor did he accompany his acceptance of the Fisher’s signal by a “corresponding alteration” of the Clement’s helm (Navigation Rule 23, Id.). Whether the Clement’s conduct, then, upon her own showing, should or should not, as matter of law, be tested alone by either or both of these rules, it is certain that her justification in omitting to observe them is essentially dependent upon the opinion of a master who admits the passing arrangement was dangerous. Necessarily the value of an opinion of this master concerning the need of observance or not of these rules would be further affected by the combined speed of the vessels. The speed at which the Fisher was then sailing was about 10 miles an hour. The normal speed of the Clement was the same, and, besides, she had the advantage of the current of the river which was about 1^ miles an hour; -but this is to be qualified by the statement of her master that she had been checked down to slow speed, whatever that may in truth mean. Clearly the precautions, if they can be called such, which the navigating officers of these,ships were taking to avert danger, are of unusual moment here.

It would be futile to attempt to reconcile the testimony as to what happened as the boats approached the point of collision. The steamers seem to have been kept on the courses they respectively selected upon the completion of the passing arrangement, until it was too late to maneuver effectively to escape danger. Much is said on both sides as to which vessel disrupted the situation. The master of the Fisher testifies that the Clement turned to her port suddenly, and “appeared to be crowding us.” And the master of the Clement insists that the Fisher sheered to the westward and across the range line or bow of the Clement, disclosing her green light, but quickly turned back showing her red light and attempted to cross the Clement’s bow to the eastward; that, in spite of the Clement’s effort to avoid collision, the bluff of her port bow struck the' Fisher on her port side aft of the boiler house. The testimony describing these movements of the ships shows a degree of excitement and confusion on the part of the respective navigating crews which was calculated to exaggerate alike their impressions and narrations of the situation and the events. This is seen, for example, in a sudden, though vain, effort of the master of the Fisher to sound a danger signal at the time he claims the Clement was crowding the Fisher. He says the gear fouled and only “two short toots” resulted. The Clement did not regard them as an ordinary two-blast signal, yet responded with two blasts. The confusion is further *587shown by repeated, and varying orders which were given by the masters to their helms. It would not be helpful to recite these orders, for they were occasioned by and so do not explain the sudden changes in courses ascribed to the respective vessels as before pointed out. The testimony bearing upon these features of the controversy fails to impress us with the belief that the Fisher alone swerved from her course sufficiently to bring the ships into collision. According to the Clement’s own version of the situation at the approach of the crisis, the distance between the two ships was not enough, as we view the testimony, to- admit of such unusual movements on the part of the Fisher; and an apparently disinterested and competent navigator of long experience on the Great Takes and the Detroit river testified, with unmistakable candor and much show of reason, that such movements were impossible. It is true that the designs and dimensions of the two vessels materially differ. The Fisher is 220 feet in, length with a beam of 40 feet, while the Clement is 480 feet in length with a beam of 52 feet; and, besides, the Fisher was relatively more heavily laden than the Clement. It must therefore be conceded under the testimony that the former was harder to steer steadily and more likely to sheer than the latter. Still the admitted fact of her recovery and return to the eastward through the use of her helm, shows that the Fisher was able to overcome the very conditions that are here claimed to have caused her to sheer. This is not to say, however, that the Fisher did not to some extent turn to her port side. We believe she did, and through failure attentiyely to watch her helm. We are not satisfied, however, that she at any time got beyond the control of her navigating officers. On the other hand, we do not think the Clement was suffered to turn so far to her port side as the Fisher claims; but upon the whole testimony we are led to- believe that she turned materially to the eastward rather than that the Fisher both sheered and recovered to the extent claimed. Indeed, if we rightly interpret the evidence touching the Fisher’s movements, there is no perceivable way to account for the disaster at all except through a movement of the Clement in material degree to the eastward. It is significant, too, that the collision, though its exact place cannot be located, occurred eastward of the north channel range. This tends to corroborate the testimony that the Clement was “crowding” the Fisher; and we cannot think that the Clement’s movement to the eastward is sufficiently explained by the turn it is said she made to avoid the collision.

Furthermore, both of these ships are open to criticism on account of their lookouts. At the time the passing arrangement was concluded, the Fisher’s lookout was not at his place of duty; and while the Clement had her lookout in proper position from the time the Fisher was sighted until the collision, the lookout was neither called to testify nor his absence accounted for. The Clement had the right of way at the start (Navigation Rule 24 [Comp. St. 1913, § 7934]) and insists, as we have seen, that she claimed it by calling for a starboard passing. The master of the Fisher qualifies his statement that he did not hear the two-blast signal by admitting that, under the circumstances, he may have heard but one of the blasts of the Clement. The result was, as *588already shown, that a passing arrangement was concluded which the master of the Clement regarded as dangerous. In view of the result, it cannot be said either that a starboard to starboard passing would not have avoided a collision, or that, if the Fisher’s lookout had been at his proper place, an agreement for such a passing would not have been made. The Fisher at this vital period, and the Clement from the time the ships were sighted, might as well have been without lookouts at all, for, as respects the times mentioned, we are in practical effect required to consider the case as though the ships in fact were in this plight. Surely, in view of rule 28 (Comp. St. 1913, § 7938), such conditions are not consistent with due care on the part of either ship. The Ariadne, 13 Wall. 475, 478, 20 L. Ed. 542; The George W. Roby, 111 Fed. 601, 612, 49 C. C. A. 481 (C. C. A. 6th Cir.); Robinson v. Detroit & C. Steam Nav. Co., 73 Fed. 883, 892, 20 C. C. A. 86 (C. C. A. 6th Cir.). The present question differs from the one in relation to the Ellwood’s lookout and passed upon, by this court in Great Lakes Co. v. Pittsburgh Co., 222 Fed. 862, May 4, 1915. Under the facts of that case it clearly appeared that the presence of the lookout would have made no' difference. The. circumstances here are not the same. The passing agreement there had been concluded before the withdrawal of the lookout, but during the corresponding period here the Fisher’s lookout was below; and, while the lookout claims to have heard the signals, it is manifest that his post of duty was the better place to hear the controverted signals of the Clement. Further, it is claimed by the Fisher, and not without some evidential support, that the comparison of the compasses of the Clement and their adjustment and reading with relation to the chart course took place and so engaged the attention of the master and second mate while the ship was on the Grosse Isle north channel range; and the Clement offered testimony tending to show that the Fisher could not be kept steadily upon a given course. It hardly need be said that the Clement’s lookout was in a position to have rendered material assistance in clearing up all these questions.

[2] If anything more were necessary in respect of the Clement’s conduct, it might be added that we think she violated rules 23 and 26. The master of the Clement admits that he gave no order to the ship’s helm either at the time of closing the arrangement to pass port to port or until the Fisher sounded the “two short toots” before mentioned; and we have seen that, when these sounds of the Fisher’s whistle were given, the ships were well within the zone of danger. It is insisted that rule 23 cannot be employed here to put the Clement in fault, for the reasons that neither the pleadings nor the course pursued in the court below will admit of it. The answering claimants, it is true, did not in their cross-libels distinctly allege the breach of this rule as a fault of the Clement; yet the testimony showing that the rule was not observed was offered and received without objection; and, moreover, the testimony came from the Clement’s master himself both as to the dangerous character of the passing arrangement and the failure to observe this rule. We therefore do not see how the consideration or effect of this testimony can be said to have taken the owner of the Clement by surprise either in the court below or here; and, under the liberal rules *589of practice prevailing in the admiralty courts, the failure to set up the matter in, the pleadings ought not to be allowed to work any injury to the cross-libelants, for such failure was plainly not designed. As Mr. Justice Davis said, in The Steamer Syracuse, 12 Wall. 167, 173, 20 L Ed. 382, where it was objected that the libel did not specifically charge an antecedent negligence as a fault:

“This is true, and the libel is defective on that account, but in admiralty an omission to state some 1'aeta which prove to be material, but which cannot have occasioned any surprise to the opposite party, will not be allowed to work any injury to the libelant, if the court can see there was no design on his part in omitting to state them. There is no doctrine of mere technical variance in the admiralty, and subject to the rule above stated, it is the duty of the court to extract the real case from the whole record, and decide accordingly.”

See, also, The Gazelle, 128 U. S. 474, 487, 9 Sup. Ct. 139, 32 L. Ed. 496; Davis v. Adams, 102 Fed. 520, 523, 524, 42 C. C. A. 493, and citations (C. C. A. 9th Cir.). If objection to the testimony had been taken below it is clear that the cross-libels might have been cured by amendment. Pioneer Steamship Co. v. McCann, 170 Fed. 873, 880, 96 C. C. A. 49 (C. C. A. 6th Cir.); Monongahela River Consol. C. & C. Co. v. Shinnerer, 196 Fed. 375, 384, 117 C. C. A. 193 (C. C. A. 6th Cir.). It certainly cannot be safely said that the collision might not have been avoided if the Clement had ported her helm upon closing the passing arrangement, or if her speed had been reduced to hare steerageway. In saying this, we appreciate the language of rule 23, which only exacts an alteration of the helm “whenever required,” and also of rule 26, which only requires reduction of the speed to “hare steerageway” when the ships have “approached within half a mile of each other”; and still it cannot be that a master can rightfully omit to alter the ship’s helm when he knows that the passing agreement he enters into is dangerous; and, further, this master himself says the ships were within, the half-mile limit when the agreement was closed.

It results that in our view the faults found as to the two ships were concurrent' in causing the collision, and that the damages must be divided. Since .the issues, the testimony, and the assignments relate only to the question of fault and the rulings thereon, as pointed out in the statement, it is neither necessary nor proper to pass' upon any of the questions, apart from those determined by the decree itself, that may arise between either of the shipowners and the cargo appellants or those representing death claims or the claim, for personal injuries.

The decree (except so far as it accords to the Buffalo Steamship Company the benefit of the limitation of liability under sections 4283, 4284, and 4285, with their amendments and supplements, - of the Revised Statutes) is reversed, with costs; and each cause is remanded, with direction to enter a decree in accordance with this opinion and order a reference therein to ascertain the damages.