Evergreen Marine (UK) Limited v Nautical Challenge Ltd

This is the first appeal in a collision action to come before the Supreme Court. Indeed, it is approaching 50 years since the last such appeal before the House of Lords - The Savina [1976] 2 Lloyd’s Rep 123.

The appeal raises two important questions of construction of the International Regulations for Preventing Collisions at Sea 1972 as amended (“the Collision Regulations”). Both questions relate to the application of the “crossing rules”, as set out in rules 15-17 of the Collision Regulations. These rules apply where two power-driven vessels are crossing so as to involve a risk of collision. They require the vessel which has the other on her starboard side to keep out of the way (the “give-way vessel”) whilst the other vessel is required to keep her course and speed (the “stand-on vessel”). They are of cardinal importance to the safety of navigation.

The first question which the court has been asked is whether the crossing rules are inapplicable or are to be disapplied where an outbound vessel is navigating within a narrow channel and has a vessel on a crossing course approaching the narrow channel with the intention of and in preparation for entering it. This concerns the inter-relationship between the crossing rules and the “narrow channel rules” - rule 9. Rule 9(a) requires vessels proceeding along the course of a narrow channel to keep as near to its starboard outer limit as is safe and practicable.

The second question which arises is whether it is necessary for the putative give-way vessel to be on a steady course for the crossing rules to be engaged. The “putative give-way vessel” is the vessel which, if the crossing rules apply, would be required by rule 15 to keep out of the way of the other vessel. In practical terms it is the vessel which has the other vessel on her starboard side. We will call that other vessel the “putative stand-on vessel”.

The collision which gives rise to these questions was between the appellant’s large container vessel, Ever Smart (“EVER SMART”) and the respondent’s VLCC (very large crude carrier), Alexandra 1 (“ALEXANDRA 1”). The collision took place within the pilot boarding area, just outside the dredged entrance/exit channel to the port of Jebel Ali in the United Arab Emirates (“the channel”), at 2342:22 local time (GMT+4) on 11 February 2015. (The time of the collision is referred to hereafter as “C” and times leading up to the collision as “C-x”.) EVER SMART was outbound from Jebel Ali laden with 48,564 mt of containerised general cargo and, until shortly before the collision, was proceeding along the channel. ALEXANDRA 1 was the inbound vessel, laden with 113,915 mt of condensate, and was, although not stationary, waiting in the pilot boarding area to pick up a pilot being released from EVER SMART before entering the channel. The channel is a “narrow channel” for the purpose of rule 9.

The experienced Admiralty judge, Teare J, and the Court of Appeal answered both questions in the affirmative, with the consequence that the crossing rules were not engaged at all or, if engaged, were overridden by the narrow channel rules, and therefore ALEXANDRA 1 was not under an obligation to keep out of the way of EVER SMART. On this appeal the appellant contends that this interpretation of the Collision Regulations is wrong and does not give sufficient weight to the fundamental importance of the crossing rules in preventing collisions at sea.

These questions are of real importance to mariners, and to the effectiveness of the Collision Regulations in preventing, as far as possible, collisions at sea. Bearing in mind that a narrow channel may include an appropriately narrow harbour entrance, the crossing situation to which the first question relates occurs very frequently, all round the world. As for the second question, there may be many reasons why a vessel which is moving over the ground may not be on a steady course, but nevertheless be approaching another vessel on a steady bearing and therefore with a risk of collision, as stated in rule 7(d)(i). It is a striking feature of the present case that, although ALEXANDRA 1 (which was the putative give-way vessel) was not on a steady course, she and EVER SMART were approaching each other on a steady bearing for the whole of the 23 minute period before the collision.

At all material times, weather conditions were benign with clear night skies and good visibility of 10 - 12 miles. The wind was force 3 said to be from an easterly or east-north-easterly direction. There was a 1 knot south westerly setting current. Both vessels were exhibiting masthead, side and stern lights in accordance with the Collision Regulations. Masthead lights are visible at a minimum distance of six miles and sidelights are visible at a minimum distance of three miles. ALEXANDRA 1, in addition, was displaying an all-round red light showing that she was carrying a dangerous cargo. Given the good visibility, the vessels would have been in sight of each other when they were about six miles apart (about C-23).

The channel lies on an axis of 315 degrees/135 degrees (true) about 8.5 nautical miles in length and slightly less than two cables in width. The channel is marked by lateral red and green buoys (port and starboard), from No 1 buoys at the seaward end to No 12 buoys at the outer breakwater and is dredged to 17 metres. At the seaward end is the designated pilot waiting area, shown on Admiralty Chart BA 3739 as a circular area with a one nautical mile radius. The limit of the circular area as shown on the chart is just under three cables beyond the No 1 buoys.

Both vessels were equipped with Voyage Data Recorders (“VDR”). As in most modern collision actions, the VDRs enabled an accurate reconstruction to be agreed, together with details of the vessels’ movements and orders, as well as providing audio recordings from the respective bridges. For the purpose of the trial, the parties agreed a reconstruction of the vessels’ tracks (ie movements over the ground) which was subsequently transposed onto Chart BA 3739 (with a scale). This was annexed to the judgments below and is annexed to this judgment as Annexe A.

A detailed account of the navigation of both vessels is set out in the judgment of Teare J at paras 23-34 - [2017] EWHC 453 (Admlty); [2017] 1 Lloyd’s Rep 666. For the purpose of the appeal the parties agreed a statement of facts summarising the most relevant matters, as set out below.

In terms of the navigation of ALEXANDRA 1:

In terms of the navigation of EVER SMART:

At the point of collision, ALEXANDRA l’s bow was on about the centre line of the channel, projected forward from No 1 buoys. The port bow of EVER SMART struck the starboard bow of ALEXANDRA 1 at an angle of about 40 degrees leading aft on EVER SMART. At collision the speed and course of EVER SMART over the ground was 12.4 knots and 316.0 degrees (true) and her heading was 323.9 degrees (true). The speed and course of ALEXANDRA 1 over the ground was respectively 2.4 knots and 104.4 degrees (true) and her heading was 101.2 degrees (true). The judge found that for the whole of the relevant period of 23 minutes before the collision the two vessels were approaching each other on bearings which, viewed each from the other, did not appreciably change: see para 66.

The loss and damage sustained by each vessel arising out of the collision has been determined (subject to apportionment) in the sum of US $9,308,594.71 for ALEXANDRA 1 and in the sum of US $2,531,373.71 for EVER SMART.

Teare J sat with Captain Stephen Gobbi and Captain Nigel Hope, Elder Brethren of Trinity House, as Nautical Assessors. The role of the Nautical Assessors is to provide advice as to matters of navigation and seamanship. The court is not bound by that advice, and must form its own view about those matters in the light of all the submissions received from the parties. The interpretation of the Collision Regulations is a matter of law for the judge to determine. The judge’s judgment was carefully structured to reflect this division of roles.

Under the heading “The applicability of the crossing rule”, the judge first considered the relationship between the crossing rules and the narrow channel rules. He was referred to and considered a number of authorities: The Leverington(1886) 11 PD 117, The Kaiser Wilhelm der Grosse [1907] P 36 and 259, The Treherbert [1934] P 31, The Empire Brent (1948) 81 Ll L Rep 306, The Canberra Star [1962] 1 Lloyd’s Rep 24, The Glenfalloch [1979] 1 Lloyd’s Rep 247, Kulemesin v HKSAR [2013] 16 HKCFA 195 and The Nordlake and The Seaeagle[2016] 1 Lloyd’s Rep 656.

The judge placed particular reliance on what he described as “statements of principle” made by Hewson J in The Canberra Star and by Lord Clarke in Kulemesin.

In The Canberra Star Hewson J stated at p 28:

In Kulemesin Lord Clarke stated as follows at para 225:

He considered that these “statements of principle” supported the case that the crossing rules did not apply in this case and said that he agreed with them and should follow them.

His other main reason for reaching that conclusion was that it cannot have been intended by those who drafted the Collision Regulations that there would be two sets of rules with different requirements applying at the same time as this would cause confusion and not be in the interests of safety. He accordingly concluded that “the crossing rules cannot have been intended to apply where one vessel is navigating along a narrow channel and another vessel is navigating towards that channel with a view to entering it” (para 53).

The judge also accepted an oral submission made by Ms Selvaratnam QC on behalf of ALEXANDRA 1 that she was not “on a sufficiently defined course for the crossing rules to apply”. He stated as follows at para 70:

Having concluded that the crossing rules did not apply, the judge then considered the alleged faults of each vessel.

One of the allegations of fault made against ALEXANDRA 1 was that she had approached too close to the end of the channel. This led the judge to ask the Elder Brethren the following question (paras 93-94):

In terms of fault, the judge concluded that EVER SMART was at fault in (i) breaching the narrow channel rule by not keeping to the starboard side of the narrow channel; (ii) keeping a defective radar and visual lookout and making assumptions on the basis of scanty information; and (iii) proceeding at an excessive speed, a direct consequence of her failure to keep a good lookout. ALEXANDRA 1 was at fault by failing to keep a good aural lookout, with the result that, following a misheard or misunderstood VHF conversation, she did not turn to starboard towards the channel and instead headed so as to cross the approaches to the channel.

The judge found the faults of EVER SMART to be very serious in terms of culpability and ALEXANDRA 1’s fault to be culpable to a substantial degree. In terms of relative culpability, the faults of EVER SMART were found to be much more culpable than those of ALEXANDRA 1.

In terms of causative potency, the judge did not consider that there was a marked difference in quality between the contribution which each vessel made to the fact that the collision occurred. Having regard, however, to the unsafe speed of EVER SMART, she contributed far more to the damage resulting from the collision than the very much lower (and safe) speed of ALEXANDRA 1. It followed that the causative potency of EVER SMART’s fault was greater than that of ALEXANDRA 1.

In the light of these conclusions, the judge’s conclusion on apportionment was that EVER SMART should bear 80% of the liability for the collision and ALEXANDRA 1, 20%.

The Court of Appeal also sat with Elder Brethren of Trinity House as Nautical Assessors, Rear Admiral Snelson and Captain Glass. The lead judgment was given by Gross LJ, an experienced Admiralty practitioner, with whom Lewison and Leggatt LJJ agreed - [2018] EWCA Civ 2173; [2019] 1 All ER (Comm) 303.

The Court of Appeal upheld the judge’s conclusion that the narrow channel rules applied to the exclusion of the crossing rules. Their reasons for so concluding are summarised at para 74 as follows:

This conclusion was “stress-tested” by considering a hypothetical East-West situation with the incoming vessel approaching the channel from the East so that it had to cross over to the starboard side of the channel to prepare for entry. It was EVER SMART’s case that the crossing rules must apply in such a case as otherwise there would be no rule of priority. The Elder Brethren were asked to consider what actions would be expected from a prudent mariner in accordance with the ordinary practice of seamen on the outbound and inbound vessel in such circumstances. Their answer set out various steps they considered that the prudent mariner should take to avoid any risk of collision. In relation to the outbound vessel this included keeping a sharp look out, ensuring compliance with rule 9 by staying on the starboard side of the channel, acquiring the inbound vessel as an ARPA target, consulting the onboard pilot and port control about the inbound vessel’s intentions, making early VHF contact with the other vessel and being prepared to adjust speed to resolve close quarter/collision risk if necessary. In relation to the incoming vessel this included matters such as keeping a sharp lookout, approaching via the fairway buoy, acquiring information about pilot boarding time and position, acquiring the outbound vessel as an ARPA target and keeping a close watch on the vessel’s bearing to determine the risk of collision, and identifying and making early VHF contact with the outbound vessel.

The Court of Appeal considered that this advice supported their conclusion that the crossing rules did not apply. As Gross LJ stated at para 84:

The Court of Appeal also upheld the judge’s conclusion that the ALEXANDRA 1 needed to be on a sufficiently defined course for the crossing rules to apply and rejected a challenge to his finding that she was not on such a course. Gross LJ considered that this conclusion was supported by observations made in various authorities as well as the decision of Brandon J in The Avance [1979] 1 Lloyd’s Rep 143. He reasoned as follows at para 92:

The questions of law which arise are in the following agreed terms:

Like the courts below, we have sat with Elder Brethren of Trinity House as Nautical Assessors, Captain Nigel Palmer OBE MNM and Commander Nigel Hare RN. Following the hearing the court considered that it needed the assistance of the Nautical Assessors. Following a request for, and receipt of, written submissions from the parties as to their form, three questions were asked of them by letter from the court on 9 December 2020. A prompt response was received from the Nautical Assessors in a joint letter dated 16 December, upon the content of which the court also sought and received written submissions from the parties. We will refer to the advice of the Nautical Assessors as appropriate during the course of this judgment. Although we have not considered ourselves bound to follow their advice we wish to express the court’s gratitude for their assistance.

The Collision Regulations are an International Maritime Organisation (“IMO”) 1972 Convention (“the 1972 Convention”). They are given the force of law in the United Kingdom and applied to United Kingdom ships “wherever they may be” by the Merchant Shipping (Distress Signals and Prevention of Collisions) Regulations 1996 (SI 1996/75), made by the Secretary of State for Transport under sections 85 and 86 the Merchant Shipping Act 1995. They are applicable in this case because one of the shipowners, the registered owner of EVER SMART, is a UK registered company.

As an international convention the Collision Regulations should be interpreted by reference to broad and general principles of construction rather than any narrower domestic law principles - see Stag Line Ltd v Foscolo, Mango & Co Ltd[1932] AC 328, 350, James Buchanan & Co Ltd v Babco Forwarding & Shipping (UK) Ltd[1978] AC 141, 152D-E, Fothergill v Monarch Airlines Ltd[1981] AC 251, 272E, 282A and 293C, Morris v KLM Royal Dutch Airlines[2002] UKHL 7; [2002] 2 AC 628, 656, para 78, and Gard Marine and Energy Ltd v China National Chartering Co Ltd (The Ocean Victory)[2017] UKSC 35; [2017] 1 WLR 1793, para 72.

Such general principles include the general rule of interpretation set out in article 31.1 of the Vienna Convention on the Law of Treaties 1969, which provides that:

The object and purpose of the 1972 Convention is to promote safe navigation and specifically the prevention of collisions at sea. As stated by Sheen J in The Maloja II [1993] 1 Lloyd’s Rep 48, 50-51:

The international character of the Collision Regulations and the safety of navigation mean that they must be capable of being understood and applied by mariners of all nationalities, of all types (professional and amateur), in a wide range of vessels and in worldwide waters. They should accordingly be interpreted in a practical manner so as to provide clear and readily ascertainable navigational rules capable of application by all mariners. They are meant to provide international “rules of the road”.

They should also be interpreted in a uniform manner and regard should therefore be had to how they have been interpreted by the courts of different countries. During the oral hearing of the present case we were only referred to one foreign court decision - the Hong Kong court decision in Kulemesin. Further foreign authority was referred to during subsequent written submissions, to which we will also refer in due course.

The interpretation of the crossing rules should have due regard to the well-known statement of Lord Wright in The Alcoa Rambler[1949] AC 236 (PC) at p 250 that “wherever possible” the crossing rules “ought to be applied and strictly enforced because they tend to secure safe navigation”. For the same reason Lord Wright stated (at p 250) that it had “been found advantageous” for a “wider scope to be given to the crossing rule” in cases of doubt on a strict application of the rules.

As Atkin LJ stated in The Ulrikka (1922) 13 Ll L Rep 367, 368:

Of particular relevance to the present case are the following Rules in the Collision Regulations:

The crossing rules form part of section 2, within Part B of the Collision Regulations. Part B is headed “Steering and Sailing Rules”. Section 2 is headed “Conduct of Vessels in Sight of One Another”, as is provided in rule 11. “Sight” means visual sight rather than for example visible on radar or AIS: see rule 3(k). It includes visibility at night, generally by observation of another vessel’s navigation lights, for which the Rules make detailed provision. Rule 15 is applicable only to power-driven vessels, so it may be described as one of the steering rules for power-driven vessels in sight of each other.

It will be necessary in what follows to make precise use of the words heading, course and bearing. None of them is expressly defined in the Collision Regulations. They are however words with a relatively settled meaning among mariners, as may be gathered from published nautical manuals and glossaries, and which it may be supposed that the Rules take for granted. The usage which follows has been checked with, and confirmed by, the Nautical Assessors, and is not a matter of controversy between the parties.

Sometimes heading and course are used interchangeably, as meaning the direction in which the vessel is being steered. But for present purposes we think it better (and the Nautical Assessors agree) to distinguish between them as follows. The heading of a vessel is the direction (expressed as a point or number of degrees on a compass) in which she is pointing at a particular moment in time. So, for example, a vessel is on a heading of North if a horizontal line projected from a compass on her centre point through her bow points North. Used in that sense, a vessel does not have to be moving to have a heading.

The course of a vessel is the direction, again expressed by reference to the points or degrees of a compass, in which she is moving. This may be through the water or over the ground. Course over the ground is sometimes called the course made good, so as to distinguish it from her course through the water. The judge uses course over the ground and course made good interchangeably. It is the course over the ground rather than the course through the water that matters for present purposes, as para 70 of the judge’s judgment makes clear. We use “course” in that sense. Where there is no wind or current the course of a vessel both through the water and over the ground may well be the same as her heading. She simply moves in the direction in which she is pointed. But this will not necessarily be so, as the Nautical Assessors have confirmed and the present case illustrates. Tidal stream, current, surface drift and wind, if present, will or may cause her course over the ground to be different from both her heading and her course through the water. Thus a vessel heading North in an Easterly current will be on a course over the ground which is East of North, the amount of the Easterly element being the product of the ratio between her speed (through the water) and the rate of the current. Broadly speaking, the slower the vessel’s speed and the faster the current, the greater will be the difference between her heading and her course over the ground. The significant difference between the heading and course over the ground of the slow-moving ALEXANDRA 1 in the present case is a good illustration.

Similarly, a wind may cause a vessel to make leeway, that is, to slide a little sideways through the water. This is almost invariably true of a sailing vessel (unless the wind is blowing from right astern) but can also be true of power-driven vessels such as high-sided container ships, as the judge noted had probably happened to the EVER SMART in the present case. Since wind is described (unlike current) by reference to where it is coming from, a vessel heading North may have her course deflected East by a Westerly wind. Leeway will produce a difference between a vessel’s heading and her course through the water. Sometimes leeway and current will act together to increase the difference between heading and course over the ground. Sometimes they may cancel each other out.

Course used as above may describe the direction of movement of a vessel at a particular point in time. More generally course may be used to describe the overall progress of the vessel over a period of time, which may accommodate changes in her heading and speed. As explained below a vessel may be said to maintain her course despite significant changes in heading and speed, measured at particular moments in time. Thus for example a vessel or a convoy proceeding along a zig zag pattern in wartime will nonetheless intend and achieve an overall specific course towards its destination, as will a vessel moving or “yawing” either side of an intended course in a rough sea: see The Queen Mary (1949) 82 Ll L Rep 303.

The bearing of a vessel, as that concept is used in the Collision Regulations, is quite different from her heading or her course. Bearing is an expression with various maritime meanings, but for present purposes compass bearing is the meaning which matters. It is the direction in which one vessel appears when viewed from another at a particular moment in time, expressed again in terms of the points or degrees of a compass. This is the way in which “compass bearing” is used in rule 7(d)(i). When vessel A takes a series of compass bearings of vessel B over time, and they do not appreciably change, vessel B is generally described as being on a steady bearing from vessel A. As the Nautical Assessors have confirmed, the compass bearing of one vessel from another may generally be measured with reasonable precision, by day or by night by radar and, when the vessels are in sight of each other, assessed visually by using a compass.

Compass (or absolute) bearing is to be distinguished from relative bearing. This is the direction in which one vessel appears from the other, measured (usually now in degrees) as an angle from the viewing vessel’s heading, from zero (dead ahead) to 180 (dead astern). In what follows, and in most of the reported cases, “bearing” is used as meaning compass bearing, in conformity with rule 7(d).

Generally speaking, a risk of collision exists between two vessels when they are approaching each other on a more or less steady bearing: see per Willmer LJ in The Aracelio Iglesias [1968] 2 Lloyd’s Rep 7, 10:

Rule 7(d)(i) provides that a risk of collision shall be deemed to exist if the compass bearing of an approaching vessel does not appreciably change. In this context “approaching” simply means getting nearer. It does not necessarily mean heading towards each other. For example, a vessel being overtaken may be heading directly away from the overtaking vessel, but if they are each on a steady bearing from the other there is nonetheless a risk of collision because they are getting nearer to each other. Rule 7(d)(i) does not make a steady compass bearing the only indicator of a risk of collision: see eg rule 7(a), (b) and (d)(ii). But nor is it merely a rebuttable presumption. Wherever it applies, the risk of collision must be taken to exist.

Section 2 seeks to deal comprehensively with steady bearing collision situations by three rules which together cover the whole of the ground. Rules 13 and 14 deal respectively with vessels on substantially the same or reciprocal courses. Rule 13 (the overtaking rule) provides that, in an overtaking situation, the overtaking vessel must keep clear of the vessel being overtaken. Rule 14 (the head-on rule) provides that two vessels approaching each other on reciprocal or nearly reciprocal courses so as to create a risk of collision must each turn to starboard. Each of those rules contains precise specifications which triggers its application: see rules 13(b) and 14(b). In order to make them work clearly at night, the lighting rules, and rule 21 in particular, operate precisely in tandem with those triggering specifications. Rule 13(b), which triggers the overtaking rule, applies whenever the overtaking vessel is more than 22.5 degrees abaft the beam of the overtaken vessel. The courses of the two vessels may well be crossing, but rule 13 overrides the crossing rules where it applies: see rule 13(a).

Rule 15 seeks to deal comprehensively with every other steady bearing collision situation, ie where the vessels are not overtaking or on reciprocal courses, but crossing courses. It provides as follows:

Rules 16 and 17, which apply respectively to the give-way vessel and the stand-on vessel, are not merely parts of the crossing rules. They each apply wherever the rules identify which vessel is to keep out of the way. They apply for example in an overtaking situation under rule 13, and where one vessel is under what may loosely be described as a special disability, under rule 18 (not under command, restricted in manoeuvrability, fishing or sailing), or between two sailing vessels under rule 12.

Rule 18 qualifies rule 15, in the sense that vessels under special disability are not bound to give way by the crossing rules. Rather the other vessel must keep clear. It has that effect because rule 15 is not among the exceptional rules mentioned at its commencement. The rules as to lights and shapes enable vessels protected by rule 18 to be identified by other vessels, both by day and by night: see rules 25 to 27.

Rule 16 imposes an almost unqualified obligation on the give-way vessel to take early and substantial action to keep well clear. The only express qualifications are (i) that the give-way vessel is not required to do the impossible: see rule 16 itself, and (ii) the general caveat that special circumstances may arise in which a departure from the Rules becomes necessary to avoid immediate danger: see rule 2(b). Rule 8 contains detailed guidance as to whether a give-way vessel should alter course, slow down or even stop. At para (a) it requires avoidance action to be undertaken “with due regard to the observance of good seamanship”.

By contrast, rule 17 imposes only a qualified obligation on the stand-on vessel to “keep her course and speed”. It is abrogated when it appears to the stand-on vessel that the give-way vessel is not complying with the Rules: see rule 17(a)(ii), or when action by both vessels has become necessary to avoid a collision: see rule 17(b). The abrogation of the stand-on vessel’s obligation does not relieve the give-way vessel of her obligation to keep clear: see rule 17(d). This re-inforces the almost unqualified nature of the give-way vessel’s obligation.

Nor is the stand-on vessel’s obligation to keep her course and speed necessarily an obligation strictly to maintain her precise heading, course, or even her precise speed. If the nautical manoeuvre upon which she is visibly engaged when she becomes the stand-on vessel involves altering her heading or course, or slowing down, she may do so without undermining the obligation of the give-way vessel to keep clear. She may for example be altering course or slowing down to pick up a pilot: see The Roanoke [1908] P 231. At p 242 Farwell LJ explains the purpose of the obligation of the stand-on vessel:

In The Taunton (1928) 31 Ll L Rep 119, the stand-on vessel was an old sailing ketch, struggling across the Bristol Channel in a very strong cross-current, occasionally substantially altering her heading to avoid being swept down from her destination in Cardiff Roads. She was held not to have failed to keep her course and speed. Scrutton LJ said, at p 120:

To this we would add two observations. First the “object you had in view” must be reasonably apparent to the give-way vessel, if the purpose of the obligation to keep course and speed, as explained in The Roanoke, is to have effect. Secondly, the “object … in view” must include, or take account of, the stand-on vessel’s obligation to comply with the other provisions of the Rules. This may include avoiding a collision with a third vessel, which may be approaching the stand-on vessel head-on, or complying with the narrow channel rule in rule 9(a) to keep to the starboard side of the channel, as is implicit in Scrutton LJ’s example of turning to follow a bend in a river.

Mention must now be made of rule 2 which, as a rule within Part A, is of general application. It is headed “Responsibility” and provides:

Attempt was made by the respondent to use rule 2 as the basis for justifying a complete dis-application of the crossing rules as a matter of construction, on the basis of an apparent conflict with the rules of good seamanship, or to treat good seamanship on its own as a sufficient alternative to the application of the crossing rules, in relation to both the questions before the court. We regard this approach to rule 2 as being misconceived. First, it is plain from rule 2(a) that compliance with the Rules is a first principle of good seamanship. The same priority appears in rule 8(a). As stated in Marsden & Gault on Collisions at Sea, 14th ed, para 5-103, rule 2(a) “merely reminds seamen of the adverse consequences of failure to comply with the rules or with the practice of good seamanship”. In The Queen Mary ([1949] 82 Ll L Rep 303, 341), Lord MacDermott said:

Secondly, rule 2(b) builds in an inherent flexibility to meet particular dangers and special circumstances which points away from an approach which simply disapplies a Rule as a matter of construction because, on particular facts, strict compliance may give rise to difficulties. Rule 2(b) contemplates not the disapplication of a Rule as a matter of construction, but justifies a limited departure from its requirements, and only in particular circumstances which meet the stern test of necessity to avoid immediate danger. As stated at para 5-127 in Marsden and Gault on Collisions at Sea, 14th ed (2016), citing The Concordia and Esther (1866) LR 1 A & E 93:

Returning to the place of the crossing rules within the scheme of the Rules as a whole, the analysis thus far suggests the following. First, the crossing rules lie at the heart of the scheme for avoiding collisions where two moving vessels are approaching each other on a steady bearing (other than dead ahead or astern) and are thereby at risk of collision. They apply wherever the vessels are not head-on and one is not overtaking the other. Both those exceptions are precisely defined. Since the obvious purpose of the crossing rules is to prevent collisions in such circumstances it follows that, as a matter of construction, they should not lightly be treated as inapplicable. On the contrary they should be applied unless there is some necessity to do otherwise. This was stated as a principle by Lord Wright in The Alcoa Rambler at p 250 (at a time when the crossing rules were contained in articles 19 and 21 of the then Collision Regulations):

Secondly, a perceived tension between the obligation of the stand-on vessel under the crossing rules to keep her course and speed and the obligation of that vessel to comply with some other rule is unlikely to be a good reason for treating the crossing rules as wholly disapplied. Rather that tension is more likely, and appropriately, to be resolved by treating the stand-on obligation as moulded for the purpose of permitting compliance with the other rule, leaving the give-way vessel’s obligation to keep well clear in full force. This is reinforced by rule 17(d).

Thirdly, the Rules are generally explicit about the effect of one rule upon another. If one Rule (here the crossing rules) is to be treated as ousted by another Rule (here the narrow channel rules) in the absence of any express provision to that effect, and there is none, then that ouster should be strictly limited to the minimum strictly necessary to avoid danger or uncertainty. An apparent tension with the dictates of good seamanship is no sufficient substitute.

The advent of radar, ARPA (Automatic Radar Plotting Aid) and AIS (Automatic Identification System) where fitted, operational and actually being used, greatly increases the ability of those navigating vessels to anticipate the risk of a collision. Radar was in general use on most commercial vessels, but not on many yachts, when the Collision Regulations took their current form in 1972. In its simplest form it provides a bearing of another vessel in much the same way as a compass, but also provides accurate ranging, so as reliably to inform one vessel when another is approaching. Radar may, if effectively used, also provide further information. For example, when “targeted” on a particular vessel, ARPA will, based on radar information, generate its course made good (ie over the ground) and can be set to provide a collision warning. AIS is based not on radar but on the GPS system. When in operation it transmits a vessel’s name, position, course and speed, whence it can be received by other AIS fitted vessels and, again, used to generate its range and bearing, and also its likely course, CPA (closest point of approach) and even an audible collision warning.

Although rule 5 (implicitly) and rule 7(b) (expressly) encourage the use of these technical facilities, the Rules should not be construed on the assumption that every vessel to which they apply is fitted with any of them, although almost all will have a compass (fixed or hand-held) capable of taking bearings. “Vessel” is defined in rule 3(a) in the widest possible terms, so as to include, among power-driven vessels, dinghies using outboard motors and yachts, large and small, when using their engines to proceed, whether alone or in conjunction with their sails. Thus, as has been noted in para 46 above, the Steering Rules only apply as between vessels in visual (rather than radar) sight of each other, and the basic method by which a risk of collision between approaching vessels is to be identified continues to be the repeated observation of the other vessel’s compass bearing over time, rather than, for example, the other vessel’s course or its CPA: see rule 7(d)(i).

This is not mere tradition or antiquarianism. If a generally applicable rule of construction of a Rule or Rules is to be laid down, it must be capable of being implemented by all vessels, as defined. Furthermore it can happen that one or more of these modern technical aids may not be operational or switched on. Thus reliance on these systems is not a complete substitute for visual observation and repeated compass bearings taken by using a fixed or (on small vessels) hand-bearing (ie hand-held) compass. In the present case, the AIS on ALEXANDRA 1 was inoperative possibly due to a vibration problem, while on EVER SMART the ARPA was targeted on the pilot launch, rather than on the ALEXANDRA 1, right up to the moment of collision. But both vessels were perfectly capable of taking bearings of each other, by compass or by radar.

It is in this context worth quoting verbatim two passages in the advice received from this court’s Nautical Assessors:

This question was the second proposed by the parties. This may be because it was an argument raised orally in closing submissions, and not subjected to detailed legal analysis at the trial, although it was in the Court of Appeal. We deal with it first because it questions the engagement of the crossing rules, whereas the relationship between the crossing rules and the narrow channel rules merely questions whether the crossing rules, although apparently engaged, are nonetheless overridden in particular circumstances.

As already noted (in para 7 above) there may be many reasons why a vessel which is moving over the ground may not be on a steady course. She may be manoeuvring through a crowded anchorage, or to pick up a pilot. She may be moving so slowly as not to answer precisely to her rudder. She may be altering her heading so as to cope with a strong tide, or she may be lying with her engines stopped and her rudder secured, drifting here and there under the influence of the wind. She may be waiting to enter a narrow channel, or a harbour, with her engines stopped, but nonetheless moving over the ground under the influence of wind or tide. She may simply be a yacht under power with an inattentive helmsman. The question is important because, if there is such a steady course requirement before the crossing rules are engaged, the putative give-way vessel may in any of those situations be relieved of what would otherwise be her obligation to keep well clear of the putative stand-on vessel, even though there is a deemed risk of collision. As stated in rule 7(d)(i) there is such a deemed risk of collision where two vessels are approaching each other on a compass bearing that does not appreciably change, regardless whether either of them is on a steady course.

Teare J provided a meticulously detailed description of the alterations in heading, course and speed of ALEXANDRA 1 in the relevant period before the collision which is fully backed up by the electronic records and agreed plot. They demonstrate that ALEXANDRA 1 was, throughout, moving over the ground very slowly, in a general East South Easterly direction. But she was altering her heading between 84 and 112 degrees and her course made good (ie over the ground) between 81 and 127 degrees. She was also altering her speed. He concluded that she was “waiting” for the pilot launch to approach, before entering the narrow channel. Ms Selvaratnam QC for the respondent submitted that the judge found as a fact that ALEXANDRA 1 was not on a course at all. Certainly he used some words to that effect, for example at para 70, contrasting being “on a course” with merely “waiting” as if the two were (contrary to our view and a number of authorities) the antithesis of each other. But the facts he found about her movement (to use a neutral word) are as summarised at the beginning of this paragraph and are not in dispute. Bearing in mind that rule 15 does not use the word “course”, the question whether that movement constitutes a course is a semantic question. For present purposes it is sufficient that, for the relevant period of just under half an hour before the collision, (i) ALEXANDRA 1 was moving over the ground in a generally ESE direction; (ii) she and EVER SMART were approaching each other; and (iii) they were doing so on a compass bearing (each as measurable from the other) that did not appreciably change. Since they were neither head-on nor was either overtaking the other, they were therefore “crossing so as to involve a risk of collision” within the meaning of rule 15.

The judge appears to have recognised this. At para 51 he asked why “the crossing rules do not apply to two vessels which are crossing so to involve risk of collision - so that the circumstances appear to be squarely within rule 15 of the Collision Regulations …”. At para 66 he stated:

Furthermore the question before the court (in agreed form) is not whether the give-way vessel has to be on a course, but whether she has to be on a steady course or, as the judge put it, a “sufficiently defined” course.

A practical and purposive analysis of this question needs to start with a clear appreciation that two crossing vessels may be approaching each other and remain on a steady bearing (with consequent risk of collision) without either vessel being on a steady course. This is because the change in bearing which an alteration in course would otherwise cause on its own may be cancelled out by a change in her speed, or by a change in either the course or (as here) the speed of the other vessel. As this case graphically demonstrates, a vessel on an unsteady course may still be crossing another’s course on a steady bearing, with a consequential risk of collision.

Moreover, leaving aside for the moment the assistance which some vessels may derive from radar, ARPA and AIS, the Nautical Assessors have confirmed that some changes in the heading (or course) of one vessel may not be readily apparent from a careful visual observation of the approaching vessel, particularly at night. Lights will readily disclose a change in the aspect presented by the other vessel either side of dead-ahead, when pairs of masthead lights converge and cross, and the exposed side light changes from red to green (or vice versa). A change in aspect either side of 22.5 degrees abaft the beam is also readily observable, because at that point the masthead and side lights are extinguished and replaced by the stern light (or vice versa). In such cases a change in the aspect presented by the observed vessel may give warning of a change in her course, and a lack of change may warn of a steady course. But significant changes of aspect may not easily be observable, and a lack of visible change in aspect may not be a reliable indicator of a steady course. Even in daylight, an observed change in the other vessel’s aspect may not be a reliable indicator of a change in her course. The change in her aspect may be caused or contributed to by the movement of the observing vessel. If the observed vessel is proceeding slowly in a strong or variable tidal stream and/or wind even a change in her apparent heading may not point to a change in her course. In sharp contrast, an appreciable change in the bearing of the observed vessel, or the lack of it, may readily be observed, using a compass or radar. In this context the Nautical Assessors have specifically confirmed that:

Leaving aside authority on this point, a practical and purposive construction of rule 15 (which alone dictates when the crossing rules are engaged) would not suggest that, in addition to the express requirements for engagement (two power-driven vessels, crossing, so as to involve a risk of collision) there is an additional requirement that either vessel must be on a steady course.

There are a number of reasons for this.

Rule 15 makes no mention of “course” at all, let alone of a “steady course” requirement, in relation to either vessel.

The Steering Rules are, generally, very clear and precise about the requirements for their engagement: see eg rules 13, 14 and 18, already examined at paras 56 to 59 above.

There is no reason to assume that rule 15 departs from that precise approach, merely because it is concisely expressed.

The risk of collision which is caused by the fact that two vessels are crossing does not depend upon either of them being on a steady course. They need only be on a steady bearing, as viewed from each other: see rule 7(d)(i), and the advice of the Nautical Assessors recorded at para 81 above. There may be other reasons why they may be observed to be at risk of collision which have nothing to do with the steadiness of their course: see the examples given in rule 7(d)(ii).

The steadiness of the course of the other vessel may be difficult to observe, whereas the steadiness of her bearing will not be. As the Nautical Assessors confirm, bearing can be precisely measured, over time, by frequent observation, using compass or radar or both.

To introduce by implication a steady course requirement as an additional condition creates a void in the protection provided by the crossing rules, where there is a risk of collision, but where there is then no applicable provision as to which vessel should give way, or other collision-avoidance guidance beyond good seamanship. Just as nature abhors a vacuum, so a purposive construction should abhor a void which leaves vessels at risk of collision without telling them what to do about it.

The steering rules should be simple and certain for mariners, professional and amateur, to understand and apply. Once it is observed that another vessel is approaching, on your starboard side, on a compass bearing which does not appreciably change, that should be enough to require you to give way, ie to take early and substantial action to keep well clear.

Ms Selvaratnam submitted that the “steady course” requirement was to be fitted into the language of rule 15 because, absent a steady course, the two vessels would not be “crossing”. But this cannot be so. It is contradicted by the very facts of this case, in which, as the judge found, the two vessels were crossing, with a risk of collision which actually occurred, even though one was not on a steady course or speed, and the other was not proceeding at a steady speed.

Ms Selvaratnam also submitted that any apparent void in protection arising from two vessels being at risk of collision without any of the main steering rules (13, 14 or 15) applying is filled adequately by the requirement in rule 2 to use good seamanship. We do not agree. Save in a head-on situation, where neither vessel has right of way over the other, it is inherently safer for two vessels at risk of collision to know which must keep clear of the other, than for each to have to take seamanlike but otherwise unspecified avoiding action without knowing what the other vessel is likely to do. Even in a head-on situation rule 14 makes clear express provision for what each vessel must do, namely turn to starboard.

We therefore turn to examine the authorities. The most important decision for this purpose is The Alcoa Rambler[1949] AC 236, a decision of the Privy Council on appeal from the Supreme Court of Canada. It was the sole authority relied upon by the judge for the conclusion that the crossing rules did not apply because ALEXANDRA 1 was not on a steady or, as he put it, “sufficiently defined” course to engage rule 15. But it is necessary to start earlier, with The Ada and the Sappho (1873) 2 Asp (NS) 4. The two named vessels were both inbound to the Humber, roughly at right angles to each other, with the Sappho on the Ada’s starboard bow. They were each, to the knowledge of the other, intending to stop to pick up pilots at the same point at the river-mouth, where the pilot vessel was moored. The crossing rules were held to apply, so that the Ada should have given way to the Sappho. Giving the judgment of the Board, Sir J W Colvile said at p 5:

In The Broomfield (1906) 10 Asp MLC 194 the steam trawler Lucania was lying with her engine stopped and her helm lashed, heading “more or less North”, waiting for the tide, and being pushed along very slightly by the prevailing SW wind, (presumably in a generally North Easterly direction) so that she was “slightly crossing” the Broomfield, which was approaching her on her starboard side, on a steady bearing. Sir Gorell Barnes P and Bargrave Deane J held that the crossing rules applied. It was the duty of the Lucania to start her engine, unlash her helm and keep well clear.

The Ada and the Sappho was followed and applied by the Privy Council in SS Albano v Allan Line Steamship Co Ltd[1907] AC 193. It was another case of two vessels proceeding to the same spot to pick up a pilot, in this case outside Halifax, Nova Scotia. They were again each approaching the other at right angles, with the Albano on the starboard side of the Parisian. The Parisian got there first and had almost stopped by the time of the collision. The Albano, although by then going full astern, hit the side of the Parisian, even though the latter had, at the very last moment, put her engines full ahead to try and cross the Albano’s approaching bow. Reversing the courts below, the Privy Council held that the crossing rules applied, notwithstanding that the Parisian was by the time of the collision almost stationary, waiting for her pilot, so that the Parisian should have kept clear. In passing Sir Gorell Barnes, giving the judgment of the Board, cited with approval the following passage from the judgment of Lord St Helier in The Pekin[1897] AC 532, 536-537:

In that case the Pekin, the putative stand-on vessel, was accused of failing to keep her course on the basis that the crossing rules applied. Both vessels were navigating on turning courses, around the bend in a river, in opposite directions. The crossing rules were held inapplicable not because the narrow channel rules applied, but because a projection of the necessarily curving courses of the two vessels would have kept them clear of each other. They were not crossing with a risk of collision.

The only factual similarity between The Albano and TheAlcoa Rambler is that both collisions took place near Halifax. In the latter case the collision occurred in the fairway of the Bedford Basin, part of Halifax Harbour, during the Second World War, when Halifax was being used as an assembly point for transatlantic convoys. The Basin is a large, deep, oval stretch of sheltered water several miles long, lying roughly North South, with an exit at the South Eastern end, leading to the Atlantic. The fairway(which does not appear to have been a narrow channel)ran from the NNW to the SSE to the exit. There were heavily congested anchorages to the West of the fairway, and a de-gaussing range (where the magnetic signature of vessels could be tested and corrected to minimise exposure to magnetic mines) to the East of the fairway. The Rambler, working as an ammunition ship and, together with escort launches, flying flags warning other vessels to keep clear, was proceeding to sea down the fairway. The Norefjord was seeking to proceed from the congested anchorage to the de-gaussing range, taking what the Board regarded as a thoroughly dangerous short cut across the fairway, on a constantly curving course, her hull hidden until the last moment from the Rambler by the numerous other ships in the anchorage through which the Norefjord was weaving her way. The Rambler had the Norefjord on her starboard side throughout, so the Norefjord was the putative stand-on vessel if the crossing rules applied. There was no opportunity for those navigating the Rambler to take bearings of the Norefjord. When her mastheads first became visible, the Rambler stopped engines, although at that stage it could not be seen whether the Norefjord was making for the exit, not involving crossing the Rambler’s course, or crossing the fairway, which would or might. As soon as the Norefjord’s hull became visible the Rambler went full astern. But this was only one and a half minutes before the collision.

The key conclusion of the Board (for present purposes) is that it was not possible for the Rambler to discern that the Norefjord was on a crossing course, sooner than she actually did. By the time that this became apparent, less than two minutes before the collision, it was too late for the remedy provided by the crossing rules to be of any use, and in any event the Rambler did all she then could to avoid the collision by going full astern. Meanwhile the Norefjord was making the Rambler’s task of ascertaining her intentions even more difficult by apparently turning to port while sounding one short blast (the signal for turning to starboard). Worse still, the Norefjord did nothing to avoid or even minimise the impending collision until the very last moment.

Lord Wright acknowledged, at pp 249-250, by reference to The Otranto[1931] AC 194, that the ordinary means whereby a vessel may ascertain whether another is crossing so as to involve a risk of collision is by taking repeated compass bearings of the other vessel. At p 250 he referred to the “critical test for inferring that the vessel is on a course” as being “that her bearing does not alter”. By “course” he is here clearly meaning a crossing course. Whether the bearing does not alter, depends, he said, upon being able to keep the other vessel under observation for a sufficient time. He continued:

There are however some dicta in Lord Wright’s judgment which have since been interpreted as laying down a general principle that, in all cases, the crossing rules are not engaged unless the putative stand-on vessel is on a steady course. If by this is meant a steadily crossing course then it is unexceptionable. An approaching vessel (which is not dead ahead or astern) will be on a crossing course. If her bearing does not appreciably change (over time), as rule 7(d)(i) makes clear, she will necessarily be on a steadily crossing course, involving a risk of collision. At p 248 Lord Wright said this:

At p 249 he said:

We do not therefore interpret this passage as meaning that, in a steady bearing case (like the present but unlike The Alcoa Rambler), there is an additional requirement to ascertain whether the approaching vessel is also on a steady heading or course. Rather we think that Lord Wright meant that a steady bearing observation was enough to show that the two vessels were on a crossing course so as to involve a risk of collision. If it were otherwise it would mean that even though two vessels were on an ascertainable crossing course involving a risk of collision, as shown by a steady bearing observation, the paradigm case for the application of the crossing rules, those rules would not apply because the approaching vessel’s actual course happened not to be “steady”, a consideration of no independent relevance to the risk of collision.

Then at p 250 he continued:

This passage, read as a whole, cannot be taken as imposing an additional steady course condition, before the crossing rules can be engaged, where the two vessels are approaching each other on an observable steady bearing so as to be on a steadily crossing course. He described the first of his two conditions as being that the two vessels be “on a course”, by which he meant a crossing course. The second condition was that the crossing course be ascertainable by the give way vessel. In other words what must be ascertainable is that the other vessel is on a course and that that course is a crossing course. These are the necessary conditions which must exist. He made these remarks about the constantly curving course of the Norefjord, as hypothetically observable from the Rambler. It had never been established in that case that the two vessels were in fact on a steady bearing at all, even if their bearings from each other could have been observed over time.

Pausing there, the combined effect of those authorities may be stated thus:

The obligation on a give-way vessel to keep well clear, imposed by the crossing rules, applies wherever it is reasonably apparent to those navigating the vessel which has the other on her starboard side that the two vessels, not being head-on or overtaking, are crossing so as to involve a risk of collision (we will call that, for short, “a crossing situation”). This is what rule 15 expressly provides.

Where the two vessels can observe each other on the open sea for a sufficient time, a crossing situation will usually be established if they are approaching each other (other than dead ahead or astern) on a compass bearing which does not appreciably change over time: ie a steady bearing: see rule 7(d)(i), The Aracelio Iglesias, The Otranto and The Alcoa Rambler.

There may be other reasons why it is reasonably apparent that the vessels are in a crossing situation: eg if both are known to each other to be proceeding to the same spot, and likely to arrive there at the same or almost the same time: see The Ada and the Sappho and The Albano. Rule 7(d)(ii) gives an express warning to this effect, and it is implicit in rule 7(a).

Where a crossing situation is thus apparent for whatever reason, there is no separate rule which disables the crossing rules merely because the putative stand-on vessel is not on a steady course. The requirement for the stand-on vessel to keep her course and speed is not a condition for the engagement of the crossing rules, but a qualified obligation imposed on her once the crossing rules are engaged.

This is a fortiori in relation to the course of the give-way vessel. She must keep well clear, even if almost stationary: see The Albano and The Broomfield. She has no “keep her course” obligation, which is imposed on the stand-on vessel once the crossing rules are engaged, so as to enable the give-way vessel to keep clear: see The Roanoke [1908] P 231.

A vessel is not freed from the effect of the crossing rules merely because she is “waiting”, provided at least that she is moving across the ground to some extent. She may be waiting for a pilot, as in The Albano, or waiting for the tide, as in The Broomfield. She may, as in the present case, be waiting both for a pilot and to enter a narrow channel or a harbour. But if she is moving then she may be crossing another vessel so as to involve risk of collision. If so (as here) the crossing rules apply.

But where the taking of bearings over time is not possible, and in particular in crowded waters where the vessels’ view of each other is impaired, then it may well be that curves or irregularities in the course of the putative stand-on vessel will make it impossible for the putative give-way vessel to ascertain that she is in a crossing situation with the approaching vessel. That is what happened in TheAlcoa Rambler. Since the working of the crossing rules depends upon how the conduct of each vessel is reasonably observable by the other, they are not engaged in such circumstances. Conversely, even if bearing observations are impossible, the fact that the stand-on vessel is on a steady course may tend to show that the crossing rules are engaged and, if they are, such a course will enable the give-way vessel more easily to keep clear: see eg The Dona Myrto [1959] 1 Lloyd’s Rep 203, 212 per Karminski J.

Both the Court of Appeal and the respondent placed reliance on two more recent cases, The Savina [1976] 2 Lloyd’s Rep 123 and The Avance [1979] 1 Lloyd’s Rep 143. In The Savina, the leading speech in the House of Lords was given by Lord Simon. The collision involved two large tankers, the Savina and the Forest Hill. From C-11 the Forest Hill was on a heading 340 degrees (true) altering to 350 degrees at C-8 and maintaining that course until collision. The Savina had been aligned on a stationary heading of 080 degrees by a tug by C-15, and thereafter moved her engines to slow ahead, staying on that course until collision, although she went full astern at C-6. The Forest Hill was the putative stand-on vessel. Brandon J had held that a crossing situation applied from C-8 but not before, both because the Forest Hill was not on a definite course before then and because it was only from then that the ships’ crossing courses gave rise to a risk of collision: see per Lord Simon at p 129 (lhc). The Court of Appeal agreed. That finding does not appear to have been directly challenged in the House of Lords. Rather it was argued that if it was apparent to the putative give-way vessel (the Savina) that a potential crossing situation (with a risk of collision) might develop in the future, then it was incumbent on her to take early action to not to let that situation arise, even before the crossing rules were strictly engaged: see p 132 (lhc).

Lord Simon rejected that submission, which he describes as a “foreshadowing” of the obligations imposed by a crossing situation, so as to apply to any vessel with a red light on her starboard bow, and inconsistent with TheAlcoa Rambler. He said:

We would readily agree that there is no duty imposed by the crossing rules at a time before they are first engaged by the existence of a crossing situation: ie a crossing giving rise to a risk of collision. As Lord Simon said, until then the two vessels must observe the rules of good seamanship. We would also readily agree that both vessels must be on a crossing course, and, where rule 7(d)(i) applies, by definition this must be a steadily crossing course since it depends on the bearing not appreciably changing (over time). It may well be that this is what Lord Simon meant by “a steady course involving a risk of collision”. If, however, he is suggesting that the crossing rules can never apply until both vessels are on a steady course, we would respectfully disagree and, for reasons already given, it is not supported by The Alcoa Rambler, as properly interpreted. Nor was such a conclusion about its extent necessary to Lord Simon’s rejection of the argument with which he was dealing. None of the cases leading up to The Alcoa Rambler (and relied upon by Lord Wright) was cited, and the time when the crossing rules were engaged in The Savina was not in dispute before the House of Lords. We have explained above why, in our opinion, The Alcoa Rambler establishes no such general proposition and why, apart from authority, it would not be a correct interpretation of the Rules. TheAlcoa Rambler says nothing about the requirements of the course of the putative give-way vessel, and even its dicta about the course of the putative stand-on vessel were heavily conditioned by the facts of the case, and in particular the limitations on visual observation and bearings caused by the crowded anchorage. The Savina was not a case in which it was found that the two vessels were in fact on a crossing course, or a steady bearing, or that there was a risk of collision, before C-8.

Although the facts of The Avance were superficially similar to those of the present case, it adds little to the weight of authority. The Avance was leaving Dakar harbour at night. The Bambara was waiting outside, to avail herself, going in, of the services of the Avance’s pilot. The Avance had the Bambara on her port side. They collided just outside the harbour while the pilot was transferring between the two vessels. The main issue in the case was purely factual: whether the collision had been caused by the Avance turning unexpectedly to port, as alleged by the Bambara, or by the Bambara unexpectedly starting at a late stage to move across her bow, as alleged by the Avance. Brandon J resolved that issue in favour of the Avance so that, on any view as to the law, the Bambara was primarily to blame. But the Bambara also alleged fault by the Avance. One count was that, as the stand-on vessel under the crossing rules, the Avance should have sounded a warning signal. This depended upon whether she was obliged to hold her course. Brandon J said that the crossing rules did not apply because the Bambara was not on a settled course at the time, nor was it possible for those on board the Avance to appreciate that she was, and he referred without further elucidation to The Alcoa Rambler as sufficient authority. It is, we think, the only prior English case in which the crossing rules have been excluded because the give-way vessel was not on a settled course. None of the other relevant authorities was referred to by the judge or, probably, drawn to his attention, and there is no analysis of the question why a case about the course of the putative stand-on vessel (The Alcoa Rambler) should be decisive or even relevant to the course of the putative give-way vessel. Nor is it clear that, before her late and unexpected move forward, the Bambara was moving at all, nor whether, if she was, she could reasonably have been perceived to have been on a steady bearing.

In written submissions following the hearing the respondent relied upon two further English authorities: The Contship Success [1998] 2 Lloyd’s Rep 488 and The Sestriere [1976] 1 Lloyd’s Rep 125, as illustrative of a supposed steady course requirement as the precondition for the application of the crossing rules, based upon The Alcoa Rambler. In the former case, the two vessels had been for some time on reciprocal (not crossing) courses which, had they continued, would have left them passing each other safely, starboard to starboard. But at a very late stage before the collision the Contship Success began a turn to starboard, in a series of 10 degree steps, across the course of the approaching Selat Arjuna, eventually ramming and sinking her. The Contship Success claimed that as a result of her unexpected turn she then became the stand-on vessel under the crossing rules, so that the Selat Arjuna should have kept clear of her. Applying The Alcoa Rambler the deputy judge held that those 10 degree turns had not been sufficiently clear and unequivocal to bring the Contship Success onto a new and (for the first time) crossing course, sufficient for the crossing rules to apply.

We have no doubt that the judge was right about that. It was a conventional application of the principle that the Rules need to be applied by reference to what reasonably appears to those navigating one vessel to be being done on the other vessel. The very late alteration of course by the Contship Success did not permit any evaluation of the risk of collision by taking compass bearings. It adds nothing to the analysis of a case (such as the present) when two vessels were crossing each other’s courses, on a steady bearing giving rise to a risk of collision, for 23 minutes before the collision.

In The Sestriere, two vessels, the Alonso and the Sestriere, had been proceeding to sea from Buenos Aires, both with pilots on board, on broadly parallel courses. The Alonso was to starboard of the Sestriere and well ahead of her. She then turned to port across the course of the Sestriere and slowed down to drop her pilot. The Sestriere carried on regardless and, after both vessels made inadequate last-minute efforts to avoid each other, they collided at very slow speed. Brandon J decided that the crossing rules did not apply because the Alonso had not settled on a crossing course. She could have turned in any direction while, or after, dropping her pilot. But he concluded that the question whether the crossing rules applied did not matter, because the Sestriere should in any event have kept clear of her. He referred to The Alcoa Rambler, as authority for the requirement that the putative stand-on vessel needed to be on a “sufficiently defined course” apparent to the other vessel. He made no reference to other authorities or to the analogous cases of The Ada and the Sappho and The Albano, which are both inimical to the notion thata vessel dropping or picking up a pilot is not on a sufficiently defined course to trigger the crossing rules. In The Albano the vessel picking up her pilot was the putative give-way vessel, whereas in The Sestriere she was the putative stand-on vessel. However that may be, we doubt whether Brandon J was right to say that the crossing rules did not apply. But the case adds nothing by way of analysis to what was decided by The Alcoa Rambler.

Also in written submissions after the hearing, the respondent referred to Trinidad Corp v Keiyoh Maru 845 F 2d 818, 1989 AMC 627 (9th Cir 1988), a decision of the US Court of Appeals, 9th Circuit. This cited and purported to apply dicta of Judge Learned Hand in Commonwealth and Dominion Line v United States 20 F 2d 729 (1927) - (CA Second Circuit). He said, at p 731:

The Court of Appeal in the present case acknowledged that most of the relevant “steady course” authorities concerned the course of the stand-on vessel, but they considered that the stand-on vessel must be able to appreciate the need to keep her course, and comply with the other obligations in rule 17(a)(ii) and (b), and that this depended upon the give-way vessel also being on a steady course. Gross LJ said, at para 92, that:

Attractive though that catchy phrase may be, we respectfully disagree with the conclusion which the Court of Appeal drew from it. Of course it takes two to cross, because the concept of crossing implies that there must be two vessels involved. It may mean that they must both be moving over the ground although, if it does, it leaves a curious lacuna in the Steering Rules, since there is no rule which requires a moving vessel to keep clear of another vessel under way, just because she has no way on. But it will be a very rare case when the bearing of a stationary vessel from a moving vessel does not change, unless the stationary vessel is dead ahead. Be that as it may, if two moving vessels are (as the Court of Appeal put it) “converging with a risk of collision”, then they must either be head-on, overtaking or crossing. If they are, as here, each on a bearing from the other which does not appreciably change (over time), and there is nothing about the circumstances to prevent each from taking visual compass bearings of the other (or from observing the bearing of the other on radar), then providing that they are getting closer to each other they are in fact, and in any event are deemed to be, at risk of collision: see rule 7(d)(i). If the vessels are converging head-on, then rule 14 applies. It they are not head-on but crossing, then the crossing rules apply unless one is overtaking the other: see rule 13. This is of course subject to any special rule, like rule 18, which affords some other “keep clear” priority, and is or may be affected by the narrow channel rules, the subject of the first question before the court.

From time to time the phrases “steady course”, “settled course” or “sufficiently defined course” have been used in this context: see eg The Sestriere per Brandon J at [1976] 1 Lloyd’s Rep 125, 130. In the present case the judge used the latter in his conclusion in para 71, and “sufficiently constant heading” in para 70. They beg the question: sufficient for what purpose? That is a question of law. In our view the course need only be sufficient to show to the other vessel that they are crossing so as to involve a risk of collision. Where the general direction (ie course) of a vessel is such as to place her on a steady bearing with another vessel, then her course is sufficiently constant or defined for that purpose, unless she is plainly embarked on a course (such as turning around a bend in a river) which will eventually avoid that result: see eg The Pekin. The judge described ALEXANDRA 1 as making “progress in a broadly east south easterly direction”, sufficient to place her on a steady bearing with EVER SMART for over 20 minutes before the collision. That must, by definition, have been a sufficiently constant or defined course for the purpose of engagement of the crossing rules, provided (as has not been challenged) that this was readily to be observed from each vessel, keeping a seamanlike look-out.

Ms Selvaratnam’s main additional submission was that, on the facts of this case, even though the two vessels were approaching each other on a steady bearing, other than head-on, they could not be perceived by the other to be crossing because ALEXANDRA 1’s erratic course left open the possibility that she might be about to make a smart turn to starboard into the narrow channel, and thereby pass clear of EVER SMART, port to port. There are a number of objections to that argument. First, the mere possibility that ALEXANDRA 1 might make a late turn to starboard was insufficient to displace the risk of collision arising from the existence of a steady bearing between the two vessels for the whole of the 23 minutes before the collision. A risk of collision is not displaced by a possibility that the collision may not happen. Secondly, where two vessels are approaching each other on a steady bearing, the risk is deemed to exist: rule 7(d)(i). Thirdly, rule 7(a) provides that if there is any doubt as to the risk of collision, then it shall be deemed to exist. Fourthly, if there is no answer to the existence in this case of a risk of collision, it makes no sense to disapply the crossing rules on a theory that the two vessels were not crossing. The risk of collision was precisely because the two vessels were crossing, as was demonstrated by their readily observable steady bearing from each other over a considerable time, which was plainly not head-on. Finally, a late turn to starboard by ALEXANDRA 1 would not have taken her clear of EVER SMART, because the latter was, to use a time-honoured phrase, hogging the centre of the channel, and showing no sign of moving over to her starboard side.

We would therefore answer Question 2 by holding that, if two vessels, both moving over the ground, are crossing so as to involve risk of collision, the engagement of the crossing rules is not dependent upon the give-way vessel being on a steady course. If it is reasonably apparent to those navigating the two vessels that they are approaching each other on a steady bearing (over time) which is other than head-on, then they are indeed both crossing, and crossing so as to involve a risk of collision, even if the give-way vessel is on an erratic course. In such a case, unless the overtaking rule applies, the crossing rules will apply.

Although it is not an issue which arises in this case, for the reasons set out above we also consider that the stand-on vessel need not be on a steady course either, even though, once the crossing rules are engaged, she must then keep her course and speed. It does not follow that she should already have been on a steady course, or speed, before the crossing rules could become engaged. Nor does the word “keep” as part of her obligation imply that the stand-on vessel must be assumed already to have been on a steady course. Rule 17, which imposes the “keep course and speed” obligation applies not merely where the crossing rules apply. It applies to every situation where one vessel is identified as the stand-on vessel in relation to another. Thus it applies to a sailing vessel, from which a power driven vessel must always keep clear except when being overtaken, under rule 18(a)(iv). It applies between two sailing vessels under rule 12, and to the vessel being overtaken, under rule 13. But in none of those situations does the engagement of the relevant rule depend upon the stand-on vessel already having been on a steady course. In all those cases “keep” means maintain the course she happened to be on when the relevant Rule became engaged. There is no reason why “keep” should mean something different when rule 17 is triggered by rule 15. And the obligation, once rule 17 is engaged, is to keep both course and speed. But there is no requirement for steady speed before rule 15 (which triggers rule 17) is engaged.

Our conclusion that the putative stand-on vessel need not be on a steady course for the crossing rules to apply runs counter to what appeared to have been a concession to the contrary made by the appellant during the hearing. At the least, the appellant stated that such a requirement did not need to be challenged, since the EVER SMART as putative stand-on vessel was on a steady course throughout. Nonetheless a conclusion to the contrary seemed to us (provisionally at least) logically to follow from the main thrust of the appellant’s case. Accordingly we invited written submissions after the hearing about whether such a concession had been made, whether it was wrongly made and whether the appellant should be permitted to depart from it.

Having received those submissions we decided that we should permit the appellant to depart from that concession. As will appear we do consider that it was wrongly made. It is right to permit the appellant to depart from it for the following reasons:

The issue to which it related is a pure question of law, namely a question of construction of the Collision Regulations.

Provided therefore that the respondent was given a fair opportunity to respond, which it has been, no prejudice would be suffered by the respondent in the court treating it as a live matter in issue.

The question whether the stand-on vessel need be on a steady course is intimately bound up with the same question in relation to the give-way vessel, which has been directly in issue throughout.

The question is one of real importance to mariners generally, even if it may only be decisive of liability in a small number of cases and is not in the present case. This is because those navigating the putative give-way vessel need to know, when they see that another vessel is approaching on a steady bearing on their vessel’s starboard side, whether a further and potentially difficult analysis needs to be made as to whether she is also on a steady course, before a decision to alter course and or speed so as to keep well clear can properly be made. In their advice to this court, the Nautical Assessors were firmly of the view that, from a seamanship perspective, no such additional requirement should be imposed. They said:

Applied to the facts of this case, the analysis must be as follows:

Both ALEXANDRA 1 and EVER SMART were moving over the ground throughout the relevant 23 minutes before the collision, even though ALEXANDRA 1 was proceeding very slowly.

Both vessels were visible to each other throughout, and their bearings from each other readily measurable, both by compass and by radar.

Both could also observe they were approaching each other, both visually and by radar ranging.

ALEXANDRA 1 had EVER SMART on her starboard side throughout.

Both vessels were throughout that period in fact on bearings from each other which did not appreciably change, and this was readily observable from each vessel. Those (reciprocal) steady bearings were not dead ahead, nor did the overtaking rule apply.

Both vessels were therefore crossing within the meaning of rule 15 and, because they were approaching each other on a steady bearing, deemed to be, and in fact, crossing so as to involve risk of collision, so as to engage the crossing rules.

ALEXANDRA 1 was not on a steady course, or speed. But she was proceeding in a generally ESE direction. EVER SMART was on a steady course but not a steady speed. The changes in the course and speed of ALEXANDRA 1, coupled with the changes in the speed of EVER SMART, were self-cancelling, in the sense that, collectively, they produced no appreciable change in the compass bearing of each vessel viewed from the other.

Therefore, subject only to the effect of the narrow channel rules, the crossing rules applied to both vessels. ALEXANDRA 1 was the give-way vessel and EVER SMART was the stand-on vessel. ALEXANDRA 1 should therefore have kept well clear of EVER SMART.

The crossing rules are expressed in general and unqualified terms but it is common ground and well established that they are inapplicable where two vessels are approaching each other in a narrow channel, one vessel proceeding in one direction along it, the other proceeding along it in the other direction, even where they appear to be in a crossing situation.

This principle was clearly stated and affirmed as having been established for many years by Willmer J in The Empire Brent at p 312:

The principle was restated in similar terms by Brandon J in The Glenfalloch at p 255:

In such circumstances the narrow channel rules ensure the safety of the vessels navigating along the course of the channel and there is no need or room for application of the crossing rules. That they might impose conflicting requirements in such circumstances is illustrated by the example of a curving channel. Vessels approaching each other in such a channel might well appear to be in a crossing situation but safety is ensured if both vessels keep to the starboard side of the channel rather than one of the vessels keeping to the course she may be on at the time when the crossing situation arises.

As Willmer J stated in The Empire Brent the crossing rules do “have to be applied from time to time in narrow channels”, giving the example of a vessel which is crossing the channel. A case which illustrates the application of the crossing rules notwithstanding that vessels were proceeding along a narrow channel is The Leverington. In that case the Leverington was proceeding up the main channel from Cardiff docks and the Rapid was proceeding down the intersecting channel from the Roath Basin. Both were narrow channels and both vessels were proceeding on the starboard side of their respective channels. Rather than keeping her course and speed the Rapid went to starboard (“ported”) in order to cross the bows of the Leverington to get to the starboard side of the main channel and there was a collision. It was held that, although the narrow channel rules applied, the crossing rules also applied. The Rapid, as the stand-on vessel, should accordingly have kept her course and speed. Had she done so the collision would have been avoided and so she was held solely to blame for the collision.

All three judgments clearly state that even though vessels are in a narrow channel and subject to the narrow channel rule, the crossing rules may also apply (see The Leverington at p 118):

Lord Esher makes it clear that the crossing rules would apply to other cases of crossing of a narrow channel, such as where a smaller river joins a larger river. If the crossing rules apply where one narrow channel intersects with another, where both vessels are subject to the narrow channel rules, they must surely also apply where a vessel simply crosses a narrow channel, where only one vessel is subject to the narrow channel rules, as Willmer J stated in The Empire Brent. This is supported by Cockcroft and Lameijer - A Guide to the Collision Avoidance Rules, 7th ed (2012), where it is stated as follows at p 75:

Another case which illustrates the simultaneous application of both the narrow channel rules and the crossing rules is The Ashton [1905] P 21. In that case the King Stephen was coming in from the North Sea fishing grounds to Grimsby. The Ashton was outbound from Grimsby andthere was a collision at the entrance to the river Humber. Gorell Barnes J held that the waterway at the entrance of the Humber was a narrow channel, that the narrow channel rules applied and that the King Stephen was in breach of what is now rule 9(a) in not keeping to the starboard side of the channel. He further held that the crossing rules also applied and that both vessels had been in breach of them. The King Stephen was in breach of the crossing rules in failing to keep her course and speed; the Ashton was in breach in failing to keep out of the way and instead seeking to cross ahead of the King Stephen. This was therefore a case where the crossing rules applied to a collision at the entrance to a narrow channel where one vessel was proceeding along the course of the channel and the other vessel was crossing the entrance of the channel in order to proceed down it.

In Marsden & Gault para 5-302 these two cases are cited in support of the proposition that “as between vessels proceeding along a narrow channel and those approaching the narrow channel with a view to proceeding along it … where in such circumstances both rules may reasonably be complied with, obedience to both rules is incumbent upon the vessels.”

The question which then arises is in what circumstances can it be said that both rules may not reasonably be complied with, so that it is necessary that one should be displaced by the other? That is the case where two vessels are proceeding in opposite directions along the course of a narrow channel. It may also be the case where one vessel is proceeding along a narrow channel towards its exit and another vessel is approaching the entry of the channel with a view to proceeding along it. The critical issue in relation to question 1 is to determine, with clarity, and as precisely as possible, when that will be so.

The courts below regarded two authorities as determinative of that question, at least on the facts of this appeal, which they both considered indistinguishable from them. They are (in date order) The Canberra Star and Kulemesin. But we would prefer to begin the analysis with The Kaiser Wilhelm der Grosse. That concerned a collision just outside the Western entrance to Cherbourg Harbour, between the Kaiser Wilhelm der Grosse (“the Kaiser”), then one of the largest passenger steamers in the world, coming out, and the Orinoco, a smaller passenger ship, going in. Each side pleaded breaches of the crossing rules against the other, and the Orinoco also pleaded a breach by the Kaiser of the narrow channel rules, not suggesting that the two could not be applied, and complied with, at the same time. Cherbourg Harbour is defined by a large, roughly semi-circular breakwater, with two entrances in it. The Western entrance is half a mile wide, with ample room for two vessels to pass port to port, each keeping to her starboard side of the entrance. During most of the relevant time before the collision the Orinoco was carefully aligning herself to enter on her starboard side. She was doing 12 knots until three minutes before the collision, slowing to half speed ahead to pass between the breakwaters close to her starboard side, having embarked her French pilot at Southampton. It is clear that she was not waiting to go in. One of the complaints pleaded against her was that she should have done just that, until the Kaiser was out and clear, but this was rejected.

On any view the Kaiser did everything wrong. She went through the entrance on a diagonal course from the starboard to the port side. She went too fast, increasing from 18 to 22 knots, and she did so in an attempt to cross the bows of the Orinoco, turning to port at a late stage. Her only complaint against the Orinoco which was pursued in the Court of Appeal was that she should have held her course, or stopped sooner.

The case is noteworthy mainly because it was the first to recognise that an entrance between two breakwaters may be a narrow channel. But it is also a case in which responsibility for a collision just outside a narrow channel was decided on the basis that the narrow channel rules rather than the crossing rules were applicable: see per Lord Alverstone CJ at p 263. There is ample reasoning why the entrance was a narrow channel, but little to explain why the crossing rules were thereby disapplied. But it is clear that the Kaiser was leaving, and the Orinoco shaping her course to come in, on her final approach, rather than waiting, either for a pilot, or for the Kaiser to come out.

The collision in The Canberra Star occurred at nightin a narrow, buoyed channel in the Lower Hope Reach of the River Thames. At the point of collision the channel has, looking down river, a right hand bend in it. The Canberra Star was coming up river but, as she approached the bend, she was cutting the corner and thereby straying onto her port side of the channel. The City of Lyons had started in an anchorage on her starboard side of the channel, and was moving into the channel to proceed down river on her starboard side, her engines full ahead, and having already worked up to 5 knots. She saw the red light of the Canberra Star on a steady bearing on her starboard side, cutting the corner. Having failed to alert the Canberra Star, she then turned hard to port, across the bows of the Canberra Star, which struck her despite going full astern at the very last moment. Unsurprisingly Hewson J held that both vessels were to blame, the Canberra Star for proceeding on the wrong side of the channel and the City of Lyons for turning to port rather than to starboard to avoid the impending collision.

The case is of importance for present purposes because Hewson J held that the crossing rules did not apply, under which the Canberra Star would have been the putative stand-on vessel. Rather the narrow channel rules applied. The City of Lyons was in fact already adjusting her course so as to be on her starboard side when she reached the narrow channel at the Cliffe Lower Buoy, and increasing her speed. She was in no sense waiting to enter it, nor did the Canberra Star think that she was. On the contrary, it was the major (but unsuccessful) plank in the case of the Canberra Star that the City of Lyons should have done just that, until the Canberra Star was past and clear.

Hewson J said at p 28:

The collision in Kulemesin occurred just inside the end of a narrow channel in the vicinity of Hong Kong, between the Yao Hai, which had just entered the channel from seaward, and the Neftegaz 67 (“the N67”) which was leaving it, but hogging the centre of the channel rather than keeping on her starboard side. The N67 took no action to avoid the oncoming Yao Hai, which was shaping her course so as to enter the channel on her starboard side, until a disastrous turn to port by the N67 at the last moment led to a collision in which more than half her crew lost their lives.

The main navigational issue in the case was whether the channel was a narrow channel at all. If it was, it was not alleged that the crossing rules applied: see per Lord Clarke at p 258, at para 201. But at para 217 Lord Clarke cited the above extract from the judgment of Hewson J in The Canberra Star, saying, at para 218: “That seems good sense to me”. Later, at para 225, he continued:

The area just outside the entrance of a narrow channel, where it meets the open sea, is a place where vessels may be engaged in a variety of different activities. It is convenient to identify three broad groups. Group 1 are vessels which are approaching the entrance of the channel, heading across it, on a route between start and finishing points unconnected with the narrow channel. They are approaching the entrance of the channel, but not intending or preparing to enter it at all. Group 2 are vessels which are intending to enter, and on their final approach to the entrance, adjusting their course to arrive at their starboard side of it. Examples include the Orinoco, the City of Lyons (although not at the entrance) and the Yao Hai. Group 3 are approaching vessels which are also intending and preparing to enter, but are waiting to enter rather than entering: ALEXANDRA 1 is a prime example, but the group also includes the Bambara in The Avance. They may be stationary, or moving, although still waiting to enter.

It was common ground between counsel, and correctly so, that the crossing rules apply as between a vessel leaving the channel, approaching its entrance, and a vessel in Group 1, regardless which of them had the other on her starboard side. This is because the approaching vessel in Group 1 is not preparing or intending to enter it (the respondent’s formulation), or shaping to enter it (the appellant’s formulation). It is also common ground, and rightly so, that the narrow channel rules, and not the crossing rules, apply as between the leaving vessel and a vessel in Group 2. That is because the approaching vessel is both preparing and intending to enter it, and already shaping (ie adjusting her course and speed to do so), on her final approach.

The debate in the present case concerns Group 3. The respondent submits that the “preparing and intending to enter” test, which they seek to derive from The Canberra Star and Kulemesin, applies to a vesselwaiting to enter, in particular if she is, like ALEXANDRA 1, waiting in a designated pilot boarding area, so that the crossing rules are overridden by the narrow channel rules. The appellant submits that no necessity arises for the crossing rules to be overridden until the approaching vessel is actually shaping to enter, adjusting her course and speed to arrive at the entrance on her starboard side of it, on her final approach, in accordance with rule 9(a). That is what the approaching vessels were all doing, in TheKaiser Wilhelm der Grosse, The Canberra Star and Kulemesin.

We agree with the appellant, for the following reasons. First, the principle enunciated by Lord Wright in TheAlcoa Rambler and by Atkin LJ in The Ulrikka that the crossing rules should be applied wherever they can, as between vessels which are in fact crossing so as to involve a risk of collision, means that they should not be overridden in the absence of an express stipulation, unless there is a compelling necessity to do so. There is or may be such a compelling necessity in Group 2 cases, as the facts of the above three cases demonstrate, but not in Group 3 cases.

In a Group 2 case, the necessity to disapply the crossing rules arises because, once she is shaping and adjusting her course to enter the narrow channel, the approaching vessel is already having her navigation determined by the need to be in compliance with rule 9(a) when she reaches the entrance, that is, to arrive at her starboard side of it, on a course which enables her to continue on her starboard side of the channel. By contrast, no such necessity affects the Group 3 waiting vessel, or any vessel approaching the channel intending to enter it, which has yet to shape her course to enter it on her starboard side of it. If she is the give-way vessel under the crossing rules, she can alter course or slow down so as to keep clear of the stand-on vessel as she emerges from the channel. In the present case ALEXANDRA 1 was the give-way vessel. She could have kept clear of EVER SMART by stopping, or by turning to starboard, even if that would have taken her outside and to the West of the green No 1 buoy, where the charted depth was no shallower than in the pilot waiting area, provided she continued to turn to starboard. If she is the stand-on vessel under the crossing rules, then she can simply maintain her course and speed.

In a Group 3 case, no similar necessity affects the vessel leaving the narrow channel. If she is the stand-on vessel, she can maintain her course, taking avoiding action under rule 17(a)(ii) or (b) only when it becomes apparent to her that the other vessel is not intending to give way, or when action by both vessels is required to avoid a collision. Then she can slow down or stop. She may be constrained by her draught not to turn to starboard but, if not, she can turn out of the channel to starboard. Rules 17(c) and 9(a) will generally prevent her from turning to port.

It was submitted for the respondent that, even in a Group 3 case, the putative stand-on vessel leaving the channel would be in a dilemma if proceeding in the centre or on the port side of the channel because the crossing rules would require her to maintain her course, whereas the narrow channel rules would require her to turn to starboard. But this is a false dilemma, for three reasons. First, as already explained, the obligation on the stand-on vessel does not mean that she must literally maintain her precise course. Keeping her course and speed includes doing so in compliance with any other applicable rule. Secondly, a question of construction about the relationship between two rules should be approached on the basis that they are being complied with, rather than (as here by EVER SMART) flouted. Thirdly the same false dilemma would arise in a Group 1 case (where the vessel crossing the end of the channel was the give way-vessel) even though it is common ground that the crossing rules apply.

The second reason for preferring the appellant’s case is that the test for the occasion when, of necessity, the crossing rules should be overridden must be a clear one, clear that is to those navigating both the vessels involved. Fundamental to the construction of the Rules is the need to apply them by reference to what is reasonably apparent to those navigating each vessel about the conduct of the other. On that basis of assessment, the test propounded by the appellant is the clear winner. The crossing rules are overridden only when the approaching vessel is shaping to enter the channel, adjusting her course so as to reach the entrance on the starboard side of it, on her final approach. That can be determined from the vessel leaving the channel by visual (or radar) observation of the approaching vessel’s course and speed. But the respondent’s test is only that the approaching vessel is intending and preparing to enter the narrow channel. Of course, the fact that she is approaching the entrance of the channel may readily be observed, but that is true also of vessels in Group 1, where the crossing rules apply. How is the leaving vessel to know what the approaching vessel is intending, or what preparations she may be making? Picking up a pilot before entering a river or a harbour entrance is clearly not a sufficient act of preparation to displace the crossing rules: see The Ada and the Sappho and The Albano. Merely being in a pilot boarding area cannot of itself be decisive, since vessels may be proceeding in that area for other reasons, eg because they are leaving the narrow channel, or merely passing its entrance en route to a completely different destination, as was the tug ZAKHEER BRAVO in the present case.

Thirdly, application of the appellant’s test will override the crossing rules in fewer cases than the respondent’s test, thereby adhering to the principle that the crossing rules should (where engaged) be strictly applied if they can be. In all three groups identified above, the crossing rules will only be engaged if the two vessels are in fact (and visibly to those navigating each) crossing so as to involve a risk of collision. A test which then overrides them in narrower circumstances than another competing test is, we think, clearly to be preferred.

Finally, we should address the Court of Appeal’s reliance on the advice of the Elder Brethren in relation to the East-West scenario of an incoming vessel crossing to the starboard side of the channel to prepare for entry as somehow endorsing the conclusion it had reached that the crossing rules do not apply. This was a matter for the court rather than the Elder Brethren. As stated in Marsden and Gault at para 7-85, the “function of nautical assessors is to advise the court as to matters of navigation and seamanship”. The proper construction of the Collision Regulations is a matter of law for the court to determine. The Elder Brethren were asked to state what actions they would “expect from a prudent mariner in accordance with the ordinary practice of seamen” in the context of the hypothetical East-West scenario. This involved giving advice as regards good seamanship, and required no assumption to be made as to what may or may not have been required under the crossing rules. Their advice was to the effect it should have been possible to avoid the risk of collision through prudent seamanship by the inbound and outbound vessels. That does not mean, however, that that risk could not have been avoided by the application of the crossing rules. Their advice, moreover, included that the outbound vessel should have been prepared “to adjust own ship’s speed to resolve close quarter/collision risk if necessary within the constraints of remaining in the channel and on the starboard side”. These are essentially the actions required of the outbound vessel as the give-way vessel under the crossing rules (which EVER SMART would have been in the East-West scenario). We do not therefore consider that their advice supports either the Court of Appeal’s conclusion or the respondent’s case on the appeal.

In our view the East-West scenario demonstrates why the crossing rules need to govern that crossing situation. The vessel approaching from the East will have to cross the line of the channel before she can turn to port to shape a course to get to her starboard side of it, taking her across the bows of the vessel leaving the channel. The narrow channel rules therefore do nothing to help either vessel avoid a collision, still less do they explain which vessel should keep clear of the other. If the crossing rules do not apply, then the risk of collision inherent in their approaching each other on a steady bearing is not addressed by any Rule.

We would therefore answer Question 1 in the negative. Where an outbound vessel in a narrow channel is crossing with an approaching vessel so as to involve a risk of collision, the crossing rules are not overridden by the narrow channel rules merely because the approaching vessel is intending and preparing to enter the narrow channel. The crossing rules are only overridden if and when the approaching vessel is shaping to enter, adjusting her course so as to reach the entrance on her starboard side of it, on her final approach.

The result of our negative answer to both the questions before the court is that the crossing rules did apply to ALEXANDRA 1 and EVER SMART for the whole of the relevant period of just under half an hour before their collision. We would accordingly allow the appeal.

But it by no means necessarily follows that this should result in some different apportionment of liability for the damage, or even responsibility for the collision having occurred and it was the respondent’s case that it should make no difference.

Neither of the parties has however asked this court to re-consider the apportionment of blame or liability, if minded to allow the appeal on the two questions of construction. Rather, they would prefer to have all matters of apportionment to be re-determined if necessary by the Admiralty Court. We agree that this would be the best course. Although Teare J has retired from full-time sitting, it would be appropriate for him to undertake that task, if available and willing to do so.