The question before us is whether the Towboat Seneca (“Seneca”) was partially responsible for the collision of the SS President Grant (“President Grant”) with a buoy owned by the United States at the upriver junction of the Chesapeake and Delaware Canal (“Canal”) and the Delaware River. After a trial in admiralty, the District Court concluded that both vessels were to blame for the collision and, pursuant to the admiralty rule of divided damages, allowed American President Lines, Ltd., owner of the President Grant, to recover from the Seneca and her owner, James McWilliams Blue Line, Inc., one-half of the damages sustained by the President Grant as well as one-half of the damages and statutory penalty paid the United States.1 The President Grant accepts this decision but the Seneca appeals. Seneca claims that the President Grant was on a collision course with the buoy without relation to the presence of the Seneca and that it was physically impossible for the tug to have been a factor in the accident. We believe, however, that there is sufficient evidence in the record to support the conclusion of the trial judge and therefore affirm.

*513I.

The basic facts may be stated simply enough although the parties dispute the exact sequence of events and the relative position of the two vessels. The night of July 12, 1961 was dark but clear; visibility was good. At about 22.51 hours the President Grant was exiting the eastern end of the Canal into the Delaware River.2 She was proceeding at a speed of approximately 8 knots (over the ground) bound for New York. Her Master, Leon S. Goltzer, was on the bridge but the actual “conning” was done by Captain Robert C. Maclntire, a licensed canal pilot. The Seneca, with a partly loaded oil barge secured alongside to port, had come up the Delaware River intending to transit the Canal westbound.3 At about 22.30 hours the tug was in the vicinity of Reedy Island which is south of the Canal entrance. The Master of the Seneca, Leonard J. Churchill, contacted the traffic control tug “Escort” for clearance to enter the Canal and was informed that permission would be granted after two ships completed their eastbound passage. The President Grant was the second of the two ships. After receipt of this advice, the Seneca proceeded to the north of a ships’ anchorage area which lies west of the main river channel and south of the entrance channel leading to the Canal. At about 22.45 hours she positioned herself in a southwesterly heading — as indicated by her lights — 50 to 75 feet astern of the last vessel in the anchorage. Captain Churchill cut her forward speed, apparently intending to maintain only sufficient power to keep her position clear of both the anchorage and the entrance channel. There was a light southerly wind and the tide — which was flood in a north-westerly direction at approximately 2 knots — would tend to push the Seneca toward the Canal entrance.

Captain Churchill, the tug’s master, testified that a few minutes after he arrived at the anchorage he sighted the lights of the President Grant proceeding through the Canal.4 It is not disputed that both vessels were in clear view of each other and were displaying proper running lights. Captain Churchill admitted, however, that he did not continuously observe the President Grant because he was concerned about keeping the Seneca clear of the northernmost ship in the anchorage. According to the tug’s master, the President Grant gave him “quite a berth” and cleared well to the north. He therefore had no reason to pay much attention to the President Grant’s passage; the vessels exchanged two whistles — indicating a starboard to starboard passage — and he brought the tug into the Canal at approximately 23.00 hours, unaware that anything amiss had occurred. He conceded later in his testimon, however, that the President Grant “headed more or less at me in my position and then gave me two whistles and changed to his [the President Grant’s] port.”

Captain Goltzer — who was corroborated by Pilot Maclntire — testified that his ship was 500 to 800 feet from the exit of the Canal when he first observed the Seneca. He estimated the tug was 800 yards away from his ship at the time, well off his starboard bow (40 to 50 degrees). His first impression was that the Seneca was “hove to.” But, when he was entering the Delaware River— about 4 minutes after making the sighting — he suddenly noted that she had moved from her original position and was lying directly across the exit at a distance of approximately 200 yards. He and the pilot agreed that the situation *514was in extremis. Rather than collide with the Seneca, he made a hard left rudder thereby swinging to port. He further testified that prior to turning, the President Grant was abreast of the buoy which was well clear about 100 feet off; the collision was almost instantaneous. And, according to Captain Goltzer, two whistle blasts were given for the sole purpose of informing the tug’s master that the President Grant was making a hard left rudder.

II.

The able and experienced trial judge concluded that the Seneca negligently blocked the Canal exit thereby embarrassing the President Grant’s navigation. He found it impossible, however, to assign a single cause for the tug’s error: either Captain Churchill miscalculated the President Grant’s speed and entered the Canal too soon, or he permitted the tug to drift northwards too rapidly by failing to maintain sufficient power to stem the tide. The trial judge also ruled that although the President Grant was placed in a perilous position, she was not free from fault; she failed to sound blasts of warning or take appropriate action to avert the developing dangerous situation.5 As we have already indicated, he therefore held that both vessels were to blame for the collision.

III.

The main argument for reversal advanced by the Seneca is that the testimony of its expert witness, Captain Joseph J. Keim, was improperly discounted.6 Captain Keim testified that under the prevailing conditions of wind and tide, the Seneca could not have drifted 600 yards —the distance from her original estimated position to a spot 200 yards directly ahead of the President Grant — in four minutes unless the tug was backing down on her engines. He also testified that because of the effect of the flood tide, the President Grant was sailing a course which would have brought it into collision with the buoy regardless of the Seneca’s presence. One reason given by the trial judge for discounting this testimony was that Captain Keim plotted the President Grant’s course using a speed of 5.3 knots rather than 8 knots, the speed testified to by Captain Goltzer. We agree with the Seneca that in this single respect the District Court might have fallen into error. While the import of Captain Keim’s testimony was not brought out with complete clarity, we will assume that he used the correct speed; and, that the figure of 5.3 knots was employed as an example to indicate that the assumed positions of the President Grant and the estimated elapsed time of 4 minutes taken to traverse the distance between those positions could not both be correct.7 However, the District Court discounted Captain Keim’s testimony for yet another reason; in computing the possible drift of the Seneca the witness failed to take into consideration the effect of the 6-knot wind on the tug and the partially loaded barge it was towing.

More basically, in considering the decision below we must take into account all the evidence in the record. It would distort the process of judicial review if we were to isolate artificially a single step in the reasoning of the trial judge, find fault with it and then hasten to the conclusion that his decision must be reversed, albeit his conclusions are sound and firmly grounded in the record and the balance of his findings. We deal in totalities and not in abstract indicia of decision. For, as the Supreme Court has reminded us, the “clearly erroneous” rule of civil actions is applicable to suits in admiralty and we are not free to upset *515the trial court’s findings unless we have “the definite and firm conviction that a mistake has been committed.” Guzman v. Pichirilo, 369 U.S. 698, 702, 82 S.Ct. 1095, 8 L.Ed.2d 205 (1962). See also McAllister v. United States, 348 U.S. 19, 75 S.Ct. 6, 99 L.Ed. 20 (1954). We lack a firm conviction that the trial judge’s decision was wrong.

The most salient fact to emerge from the testimony of the three witnesses who actually participated in the events leading up to the collision, is that not one of them could state with assurance the relative positions of the two ships at specified times. Again and again the words “approximate,” “estimate,” and “about,” appear in the record. And this is not surprising. Captain Churchill, who was unaware of the collision until two weeks after it occurred, had no reason to fix in his mind the precise sequence of events. As he candidly admitted, he testified to what he would ordinarily do and not what he actually remembered doing. Moreover, as noted by Captain Goltzer:

“We are dealing with large ships and confined in cramped waters at night. And the rate of occurrence at which these events take place is quite rapid.” A further cause of the imprecision that permeates the record is that the crucial 4-minute figure — the time lapse between the President Grant’s sighting the Seneca and the collision — is apparently derived from figures that were not entered in the President Grant’s log immediately following the collision but during the following watch; Captain Goltzer was too busy looking to the safety of his vessel to attend to the log.8

Under these circumstances, the most that can be said for Captain Keim’s calculations is that they were based on uncertain estimates of time, speed and position. It is hardly surprising that the trial judge preferred to credit the unequivocal testimony of the only witnesses who had cause to remember the actual events. And, Captains Goltzer and MacIntire both testified that the Seneca was almost directly in their path and that the collision with the buoy occurred only after the President Grant swung to port.9 Moreover, as we have already noted, Captain Churchill admitted that the President Grant was headed more or less directly at him and changed to port after giving two whistles. No convincing reason has been suggested why the President Grant, which was proceeding eastbound through the Canal, should have swung to the right into the anchorage — where the Seneca allegedly was still located. The far more plausible explanation is that Captain Churchill miscalculated and did not stay clear of the Canal entrance long enough. As a last gasp, the Seneca argues that the exchange of whistles merely indicates that the ships were within a half mile of each other. Griffin, On Collision 202 (1949). But this overlooks the clear testimony of Captain Goltzer, which the trial judge could properly credit, that he gave the signal to assist the tug only after the situation was in extremis. We find no merit in the Seneca’s remaining arguments.

Affirmed.