July 1971 marked the fifteenth anniversary of the Andrea Doria-Stockholm collision in which 51 persons perished. Six months after the disaster, the official inquiry was terminated when, apparently convinced by the welter of contradictions and name-callings that guilt or innocence could not clearly be ascertained, both steamship companies agreed to assume equal responsibility for the disaster. With the permission of the Federal Court, all third-party claims were settled under the terms of “limitation of liability” whereby a potential total of $85 million in claims was settled for $6 million.
On the first anniversary of the disaster, George Horne, marine editor of The New York Times, wrote: “In the morass of legal actions that followed the Stockholm-Doria crash, no impartial body ever rules on the accident in an official determination as to which, if either, of the ships was at fault and why. The suits for death and injury damages, and losses to personal effects and cargo were neatly packaged in a settlement procedure approved by the Federal District Court, in such a way that a final legal determination of liability and negligence, such as would almost naturally follow an airline, railroad, bus or automobile accident has not yet been established.” This quote of Horne’s is as applicable today as it was when it was written.
In his editorial, Horne twice refers to the “accident,” a word which has inherent connotations of the unforeseeable and unavoidable. But, before a catastrophe of this magnitude can be so labeled, at least three questions ought to be asked: What happened? How did it happen? Why did it happen?
On that fateful night of 25 July 1956, the westbound Italian luxury liner Andrea Doria was rammed by the eastbound Swedish motorship Stockholm at 11:11 p.m. The collision occurred 180 miles east of Ambrose Lightship, the main entrance to New York Harbor. The Stockholm's bow penetrated deep into the Andrea Doria's hull at a point directly under her bridge on the starboard side in what was approximately a right-angle collision. The duty officers on the Andrea Doria’s bridge said the night was foggy with visibility reduced to less than one mile. The Stockholm’s bridge watch claimed the night was clear with scattered patches of insignificant fog. Until about one minute before the collision, when her engine order telegraphs were suddenly flashed to “Full Speed Astern,” the Stockholm was running at 18.5 knots with her telegraphs set on “Full Speed Ahead.” The Andrea Doria was running at a speed reduced from 23 to 21.85 knots, with her engines operating under “Stand By” orders. This she had been doing for several hours prior to the collision because of restricted visibility. The Andrea Doria sank some ten hours after the accident while the Stockholm limped back to New York with about 75 feet of her bow demolished.
At the official inquiry the testimony of the two prime witnesses was in total conflict even to the point as to where each claimed to have first observed the other appearing out of the fog. Ernest Carstens-Johannsen, the Stockholm’s Third Officer, who was in charge of her bridge at the time of the collision, testified that he saw the Andrea Doria appear from out of the fog over the Stockholm’s port bow. At the same time, Captain Piero Calamai, of the Andrea Doria, claimed that when the Stockholm appeared from out of the fog he saw her over the Andrea Doria’s starboard bow. These were visual sightings; thus, any possibility of malfunction of the radar equipment must be ruled out.
Common sense dictates that when two ships are approaching each other on reciprocal courses, the passing must be either starboard-to-starboard or port-to-port. The testimony of the two men in charge of their respective bridges at the time of the collision contradicts this reasoning. The only possible conclusion is that one witness is correct and one witness is incorrect. There is no room for compromise. Fortunately, there exist two vital documents upon which is inscribed the full story of how the ships arrived at the collision. These documents are the Sperry Gyrocompass Course Recorder graphs which had been officially presented by both defendants and recorded by the Court as evidence.
In reconstructing collision cases, strandings, or any maritime casualty where navigation is involved, all that the U. S. Coast Guard, or Navy, requires are the Course Recorder graphs and the testified speed of the ships involved. Through the interpretation of these known factors, the events leading up to a casualty can be accurately determined. The absolute integrity of this Sperry instrument has never been questioned, let alone challenged.
Had this inquiry been conducted under the jurisdiction of the Coast Guard, or Navy, the first order of business undoubtedly would have been to produce a plot or diagram illustrating that which was scribed on the graphs. As each witness testified, his testimony then would have been compared with the graph’s evidence. In this manner the correct answers would have been apparent. But this was not done, since neither the U. S. Coast Guard nor the Navy investigated the collision which involved two foreign ships in international waters and was therefore not under their jurisdiction. Let us, then, reproduce a plot of our own.
Figure 1 has been produced from the Course Recorder graphs. The plot is to scale and has been worked back minute by minute from 11:11 p.m. to 11:04 p.m. where the actual collision sequences began.*
Andrea Doria’s Testimony
[A] to [A] Captain Calamai and his Third Officer sight glow of the Stockholm’s lights. Believing the ships to be on parallel courses, the Doria's captain maintains speed and course.
[3] to [3] Seeing that the Stockholm is turning sharply to her right—toward and into the Andrea Doria—Captain Calamai orders hard left turn and maintains speed on engines.
ⓧ The collision with the Stockholm occurs just as the Andrea Doria begins to answer her helm.
Stockholm’s Testimony
[1] to [1] Third Officer sights the Andrea Doria at a distance of 1.8 miles, bearing 20 degrees over the Stockholm’s port bow, and orders 22½-degree right turn.
[2] to [2] The Stockholm completes right turn and Third Officer checks the Doria’s lights. He then turns to answer telephone.
[3] to [3] Seconds later, Third Officer finishes telephone conversation and again checks the Doria's lights. He now finds the Doria is in a collision situation and is attempting to race across the Stockholm’s bow. He orders hard right turn and full speed astern on both engines.
ⓧ The collision occurs with the Andrea Doria after the Stockholm has turned 13 degrees.
Figure 2 is an enlarged illustration of the testimony of the Stockholm's Third Officer as lifted from Figure 1. The “S”-turn illustrated in Figure 2 is the fastest, most direct route for the Andrea Doria to get from point 2, where the Third Officer placed her one minute before the collision, to point 3, where he put her seconds later at the end of his telephone conversation. The Andrea Doria’s turning maneuver is based on Brown’s Formula, which says the diameter of a ship’s turning circle is from 7 to 9 times her length. With these facts established, the Andrea Doria would have been required to increase her speed to better than 2,500 miles per hour to maneuver herself from point 2 into the collision situation at point 3.
Neither the Swedish American Line, the owners of the Stockholm, nor its attorney, the late Charles S. Haight, agreed with the interpretation of the Andrea Doria’s graph as outlined in Figure 1. In his book about this disaster, Collision Course, Alvin Moscow informs us that they contend the Andrea Doria was at point 4 (shown on her graph) and not at point X from which point Figure 1 was produced. Moscow explains this claim in detail saying that the graph shows that the Andrea Doria executed a long three-minute turn of 110 degrees to port going from a compass heading of 275 degrees to a heading of 165 degrees. He continues by crediting Haight with the statement that the accident must have happened at the end of this long port turn.
Figure 3 is an illustration of this contention made by Haight. The Andrea Doria's reduced speed of 21.85 knots is indicated in the one-minute (X-4 Min.) period prior to X-3 Min. She would have been required to double her speed to reach this collision position. Furthermore, in no way can the condition of the Stockholm's bow after the collision be reconciled to the roll-type collision illustrated here.
Long before the Course Recorder graphs became available for study, this writer did a comprehensive study of the testimony of the official inquiry. The study eventually became the subject of an article, “There Must Have Been a Third Ship!” which was published in the July 1958 issue of the Proceedings. The outcome of this probe was given in a single paragraph that stated, “In analyzing the testimony it is plain to see that, barring perjury, there is only one possible conclusion—The original lights sighted by the Stockholm were not those of the Andrea Doria!” If the Stockholm’s navigational watch was correct in seeing lights over the Stockholm's port bow three minutes before the collision, then these lights were being displayed from a ship other than the Andrea Doria. And, in turning away from these lights, the Stockholm plunged directly into the Andrea Doria coming up on the Stockholm's starboard side. The Course Recorder graphs confirm this theory.
There are certain irrefutable facts that should convince anyone about the validity of the third ship theory. The Stockholm's Third Officer said that three minutes before the collision the Andrea Doria appeared out of the fog. A positive radar fix at that time placed her at a distance of 1.85 miles, bearing 20 degrees over the Stockholm’s port bow. The Andrea Doria, he continued, was showing her red sidelight with masthead and range lights open. At this juncture he ordered a two-point (22½-degree) turn to starboard away from the Andrea Doria. Then, three minutes later, his ship plunged halfway through the Andrea Doria’s hull, at a point directly underneath her green sidelight.
Under no circumstance could the Andrea Doria have been anywhere on the port side of the Stockholm three minutes before the collision. Furthermore, it would have been utterly impossible for anyone on the Stockholm’s bridge to have seen the Andrea Doria’s red sidelight before the collision.
Both ships have categorically denied that there was any other ship in the immediate vicinity and, stated that during the entire collision approach there never was more than one target appearing on their radar screens. Moscow, in his book, Collision Course, tells us that the lookout in the Stockholm’s crow’s nest testified that, after he had completed his telephone conversation with the Stockholm’s Third Officer one minute before the collision, he watched the Andrea Doria execute the long port turn into the collision. Both Walter Lord, author of A Night to Remember, in his review of Alvin Moscow’s documentary, published in The New York Times, and Captain Raoul de Beaudean, commanding officer of the SS Ile de France which played such a heroic role in the rescue operations, have totally discounted any possibility of a third ship being involved in this tragedy; in fact, both ridiculed the idea.
Yet, if there was no third ship, and if the Stockholm was not turning away from lights showing on her port bow three minutes before the collision, why did the Third Officer order and execute the bold 22½-degree starboard turn that ended in disaster?
A comparison of Carstens-Johannsen’s testimony with the established facts of the case, implies an error in his use of radar as an aid to navigation. The circumstantial evidence clearly indicates that the Stockholm's radar equipment was operating on the five-mile range while the Third Officer calculated his interpretations as though the equipment was operating on the 15-mile range. This error in radar range could well have been what triggered off the fatal sequence of events.
The Stockholm's Course Recorder graph shows a three-degree change in course to starboard which started at 11:06 p.m. Therefore, the order for this change in course was issued at 11:05 p.m., plus or minus a few seconds, which was six minutes before the collision. The Third Officer said that just as he ordered the slight change in course to starboard, which was to compensate for drift which he found as a result of the 11:00 p.m. radio direction bearing that he had recorded in his log book, be picked up the radar pip of the Andrea Doria at a distance of 12 miles.
This was obviously a gross error, because six minutes before the collision the Andrea Doria was only four, and not 12, miles away from the Stockholm. (See Figure 1—Stockholm’s approach at 11:05 p.m. marked “Error.”)
Moreover, at 10:53 p.m., when the ships actually were 12 miles apart, there is no change in course indicated on the Stockholm's Course Recorder graph. (See Stockholm's graph, point 10:53.) Here the graph shows the Stockholm on a steady course of 92 degrees with considerable yawing from 10:40 p.m. until 11:06 p.m. when she made the slight change in course of three degrees to starboard.
All this would seem to indicate that the Stockholm’s radar equipment was operating on the five-mile range, while the Third Officer considered it to be operating on the 15-mile range. A target four miles distant on the five-mile range setting would appear on the identical spot on the radar’s screen as one on the same bearing would have if she were 12 miles away on the 15-mile setting. (See Figure 5.) The Third Officer also said that the target was slightly to the left of the Stockholm’s heading flasher (course line indicator on screen). According to Figure 5, which has been produced from Figure 1, the radar target was actually slightly to the right of the heading flasher. The two broken lines appearing on Figure 5 illustrate how the Stockholm was yawing to the right and left of her projected course as indicated on her graph. With this constant yawing, it is plain to see how the target would jump from one side to the other of the ship’s projected course line or heading flasher.
Carstens-Johannsen also testified that the target was coming toward the Stockholm at great speed. This, he said, led him to believe that the target might be some sort of naval vessel on maneuvers. Then, three minutes before the collision, when the radar target had actually closed from a distance of four to two miles (while through his misinterpretation he thought the target had closed from 12 to six miles—thus indicating great speed), Carstens-Johannsen committed what appears to be the fatal error. It was at this moment that he ordered the 22½-degree change in course to the right which brought the Stockholm into the collision situation with the Andrea Doria (see Figure 6).
Had the Andrea Doria actually been at a distance of six miles, instead of the two miles that she was, three minutes before the collision, the Stockholm’s Third Officer’s action in ordering this 22½-degree turn to starboard would have been absolutely correct according to the rules for using radar as an aid to navigation. Here the rule is: When a radar target appears dead (or nearly dead) ahead, make an early and substantial change in course, preferably to the right, in ample time for your action to show clearly on the other ship’s radar screen. With the elimination of any other ship being present in the immediate area, this is the only other logical reason for Carstens-Johannsen to order such a bold change in course.
After the Stockholm had leveled off from the 22½-degree right turn, which was about one minute before the collision, the Third Officer said that he momentarily removed his eyes from the Andrea Doria to answer the telephone. At the conclusion of this telephone conversation, which had been from the lookout in the crow’s nest reporting the lights, he found that the ships had totally changed relative positions. Now, he said, the Andrea Doria was in a collision situation attempting to race across the Stockholm's bow (see Figure 7). From this point on, right up to the collision, the testimony, as given by both ships and the documentary evidence scribed on their graphs, dovetails in every respect. Thus, the testimony and graphs have established that the first time that Carstens-Johannsen could possibly have seen the Andrea Doria was about one minute before the collision after he had completed the 22½-degree right turn. It also establishes that the ordering of this 22½-degree right turn three minutes before the collision was the final cause of the disaster.
At the official inquiry, Carstens-Johannsen testified that he had changed the range on his radar equipment from the five-mile to the 15-mile to the 50-mile ranges from time to time during the course of his watch. When one considers that the range in radar equipment is changed with the turn of a dial, as one would change channels on a television set, it is understandable how easy it would be, in the pitch black of a darkened wheel-house, to commit such an error. This is not an infrequent kind of error; however, in the majority of cases, it is detected before any damage is done. Nevertheless, it could account for the all-too-many collisions in which ships find themselves approaching each other in the fog on starboard-to-starboard reciprocal courses which terminate in right-angle type collisions.
For those whose prime interest in accident cases is safety and education, there is no satisfaction in pointing the finger of guilt. Yet, facts are facts, and they have an eloquence of their own. Accidents don’t just happen; they are caused. By examining the causes, valuable lessons can always be learned, by and for the men who are and will be in charge of the watch on the bridges of ships.
* A 10-degree error in synchronization between the Andrea Doria’s Gyrocompass and Course Recorder was detected, proved, admitted, and explained by her captain. This error is reflected throughout the plot; i.e., the plot shows the Andrea Doria on a heading of 269 degrees at 11:05. Her graphs show a heading of 279 degrees at the same time. Also, the Stockholm’s graph shows the collision at 03:11, the Andrea Doria’s at 12:11 while the actual zone time was 23:11. This discrepancy is of no consequence, because in producing a plot, the ships are placed in their respective collision positions at the moment of impact. Then, with each horizontal line representing a period of 10 minutes, the plot can be worked back minute by minute to a point as far as is necessary to determine just how the ships entered the collision. The "note” appearing on the left side of Figure 2 calls attention to a fluctuation which could only be caused by a tremendous jolt.