England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM
(Sir Peter Webster)

		Royal Courts of Justice Strand London WC2
		10th May 1996

LORD JUSTICE SIMON BROWN: During the 18-month period from April 1990 to October 1991 Lindis Elizabeth Percy and Rachel Diana Greaves (the plaintiffs) between them were arrested over 150 times and removed from the vicinity of the Menwith Hill Station, a military communications installation in North Yorkshire used jointly by the US National Security Agency and by GCHQ personnel. The plaintiffs' activities, they say, are designed to stop abuses of power by the Ministry of Defence (MOD).

Trespass is not, of course, an arrestable offence - not, indeed, a criminal offence at all. The plaintiffs were arrested rather for alleged breaches of the HMS Forest Moor and Menwith Hill Station Byelaws 1986 (SI 1986 No.481) (the Byelaws). It is the validity, and effect of any invalidity, of these which are the issues at the heart of this appeal. In a sentence, the plaintiffs contend that the Byelaws were void for uncertainty and can never, therefore, have founded any lawful arrest.

The defendants are 66 constables of the MOD Police who carried out the arrests, together with the Chief Constable of North Yorkshire and the Attorney-General. The individual constables are each sued for wrongful arrest and false imprisonment. The Chief Constable is sued pursuant to section 48 of the Police Act 1964 in respect of the plaintiffs' various detentions by civilian police officers into whose custody the military police had passed them - further occasions of alleged false imprisonment. The Attorney General is sued in place of the Secretary of State for Defence pursuant to section 17 of the Crown Proceedings Act 1947, his liability being said to arise by way of breach of statutory duty for making allegedly defective byelaws.

On 21st July 1994 Master Miller ordered the trial of two preliminary issues: first, as to the validity or otherwise of the Byelaws; second, as to whether, if the Byelaws were invalid, the plaintiffs' claims against the constables must necessarily succeed - whether, in short, such a finding of invalidity would deny the constables any defence of lawful justification. Those issues were substantially re-formulated by Sir Peter Webster (sitting as a Deputy High Court Judge) at the trial. For present purposes, however, the precise formulations do not matter: it is sufficient to indicate merely that the judge found essentially for the plaintiffs on the first issue and for the defendants on the second. Neither side, however, is entirely happy with the judge's actual rulings on either issue and both now are either appealing or at least seeking variations of them.

With that brief introduction let me at once turn to the relevant parts of the Byelaws and the empowering legislation. The Byelaws were made by the Secretary of State for Defence under Part II of the Military Lands Act 1892. Section 14(1) of that Act provides:

"Power of Secretary of State to make byelaws as to use of land held for military purposes and securing safety of public

(1) Where any land belonging to a Secretary of State or to a volunteer corps is for the time being appropriated by or with the consent of a Secretary of State for any military purpose, a Secretary of State may make byelaws for regulating the use of the land for the purposes to which it is appropriated, and for securing the public against danger arising from that use, with power to prohibit all intrusion on the land and all obstruction of the use thereof.

Provided that no byelaws promulgated under this section shall authorise the Secretary of State to take away or prejudicially affect any right of common."

(That proviso, it may be noted, although immaterial to the present appeal, proved fatal to the Greenham Common Byelaws - see DDP v Hutchinson [1990] 2 AC 783.)

To understand the full territorial scope of section 14(1) the following additional two provisions are relevant:

"14(3) For the purposes of this section, 'land belonging to a Secretary of State' means land under the management of a Secretary of State, whether vested in Her Majesty or in the Secretary of State, or in a person as trustee for Her Majesty or the Secretary of State; and 'land belonging to a volunteer corps' means any land vested in that corps or in any person as trustee for that corps."

Section 23 provides that :

"....the expression 'land' includes any easement in or over lands....."

Section 17(1) provides:

"Notice and enforcement of byelaws

A Secretary of State, before making any byelaws under this Act, shall cause the proposed byelaws to be made known in the locality, and give an opportunity for objections being made to the same, and shall receive and consider all objections made; and when any such byelaws are made, shall cause the boundaries of the area to which the byelaws apply to be marked, and the byelaws to be published, in such a manner as appears to him necessary to make them known to all persons in the locality; and shall provide copies of the byelaws being sold at the price of [5p] for each copy to any person who desires to obtain the same."

Section 17(2) (more directly relevant to the second issue than the first) provides:

"If any person commits an offence against any byelaw under this Act, he shall be liable, on conviction before a court of summary jurisdiction, to a fine not exceeding [level 2 on the standard scale], and may be removed by any constable or officer authorised in manner provided by the byelaw from the area, whether land or water, to which the byelaw applies, and taken into custody without warrant, and brought before a court of summary jurisdiction to be dealt with according to law, and any vehicle, animal, vessel, or thing found in the area in contravention of any byelaw, may be removed by any constable or such officer as aforesaid, and on due proof of such contravention, be declared by a court of summary jurisdiction to be forfeited to Her Majesty."

As will appear, section 2 (5) of the Military Lands Act 1900 (an Act to be construed as part of the 1892 Act - see section 6) is also relevant upon the issue of certainty:

"Where an area to which byelaws under this section apply consist of any sea or tidal water, or the shore thereof, and the boundaries of the area cannot, in the opinion of the authority making the byelaws, be conveniently marked by permanent marks, those boundaries shall be described in the byelaws, and shall be deemed to be sufficiently marked within the meaning of section seventeen of the Military Lands Act 1892, if, while the area is in use for military of naval purposes, sufficient means are taken to warn the public from entering the area."

Now to the Byelaws themselves. They were made on 10th March 1986 to come into operation on 26th March 1986. Byelaw 1, under the heading "Application of Byelaws", reads:

"The areas to which these byelaws apply consist of lands belonging to the Secretary of State in the parishes of Menwith with Darley and Birstwith and Felliscliffe and Norwood and Fewston in the County of North Yorkshire, which lands are for convenience of identification shown by a thick black line on the plan annexed to these byelaws and identified as 'Plan of HMS FOREST MOOR and Menwith Hill Station" all of which is hereinafter referred to as 'the protected Areas'."

The plan annexed is clearly small-scale (it is put by the plaintiffs at about 1:25,000). It depicts two large areas (respectively Forest Moor and Menwith Hill) defined by thick black lines separated at one point by what appears to be a road. In places the areas depicted are crossed by what a legend shows to be public footpaths. A third very small square area appears just north of Menwith Hill, perhaps connected to it by a footpath.

Byelaw 2 specifies the prohibited activities which include entering the protected areas except by way of an authorised entrance, and remaining in them after being directed to leave. Byelaw 3 provides that anyone contravening or attempting to contravene Byelaw 2 shall be guilty of an offence. Byelaw 4 defines the persons "authorised to remove from the Protected Areas and to take into custody without warrant any person committing an offence against Byelaw 2" (and to remove too any objects found inside).

An explanatory note at the end contains information as to where copies of the Byelaws and plan may be inspected and obtained.

Let me next indicate something more of the history of this dispute. Having been prosecuted for a number of breaches of the Byelaws Miss Percy (and a man named John Bugg, now deceased, earlier a third plaintiff in the action) were in May 1991 acquitted of all charges by the stipendiary magistrate at Ripon. It is not necessary to relate the precise course of those proceedings; suffice it to say that the enforceability of the Byelaws was in question.

The following month Miss Greaves was twice arrested, whereupon, acting in person, she sought judicial review of her arrests. She attacked the validity of the Byelaws on a number of different grounds. Her renewed application for leave was refused by the Divisional Court (Woolf, L.J. and Pill. J.) on 26th November 1991. One short passage of the judgment (unreported, given by Pill, J.) is relevant to the present appeal:

"I do not consider it an arguable ground for declaring the Byelaws invalid that.....the map is inadequate..... . The map was not required to be a definitive marking and it is sufficient to comply with the law. The first paragraph of the Byelaws does set out sufficiently the area to be covered."

(The writ in the present action, one may note, had been issued the previous month.)

Finally, and most importantly in the history, in July 1992 there came before the Divisional Court (again Woolf, L.J. and Pill.J.) two conjoined appeals by way of case stated: one by John Bugg and Miss Greaves against the DPP in respect of their convictions under Byelaws promulgated for RAF Alconbury; the other by the DPP against the dismissal (above referred to) of the charges against Miss Percy and John Bugg under the Forest Moor and Menwith Hill Byelaws. The first appeal was allowed, the second dismissed.

The court's judgment on that occasion was given by Woolf, L.J. and is reported: [1993] QB 473. The argument had extended over five days and the great bulk of it, together with most of the judgment, was concerned with the important question, raised by both appeals, as to the circumstances in which the validity of subordinate legislation can be challenged in the course of criminal proceedings. Of altogether greater relevance to the present appeal, however, are the Divisional Court's conclusions on the self-same issue as now arises with regard to the validity of these particular Byelaws. The court there found them to be insufficiently certain and in the result "defective on their face in this regard". The most critical passages in the judgment are these at page 503:

"Are the byelaws sufficiently certain?

Byelaws such as are here under consideration which create offences must clearly state what action is required in order to commit an offence. A person who is subjected to the byelaw is, therefore, entitled to be given the necessary details to enable him to avoid contravening the byelaw. ......

In the case of the Forest Moor Byelaws, there is no description contained in the byelaws as to the boundary of the protected area. The only reference is to the lands belonging to the Secretary of State. Mr. Bishop [appearing as amicus curiae, the Ministry of Defence having declined an invitation to present argument to the court] says that this does not matter because of the obligation under section 17(1) of the Act of 1892 of the Secretary of the State to mark the boundary on the ground. However, if this submission is right, it would mean that the area subject to the byelaws would be dependent upon action being taken by the Secretary of State which he may or may not take and which would enable him to fix the boundary at will so long as he defined that boundary within his own lands.

We do not accept that this can be the situation. In our judgment, before the Secretary of State is entitled to rely on what is marked on the ground, or for that matter on some other document, there must be some reference to the marking on the ground or the other document in the byelaws themselves. This does not occur in the case of Forest Moor."

The court had earlier noted that it was common ground that the black line surrounding the protected areas on the plan would represent approximately 60 feet on the ground. Mr. Howell, Q.C., I should perhaps note, makes no such concession here.

Mr. Pleming, Q.C. for the plaintiffs not surprisingly places considerable reliance on the decision in Bugg. Plainly, as he recognises, this court is not bound by Bugg - and, of course, the MOD not having been a party to those proceedings, no question of issue estoppel arises. Having regard to the constitution of that court, however, its decision clearly commands the greatest respect.

That, understandably, was a major consideration in Sir Peter Webster's approach to the case. Put shortly, he thought it right to follow Bugg although he then went on to indicate that but for Bugg he would have decided:

"That the boundaries of the protected area are sufficiently defined by the black lines shown on the plan and by the permanent marks along the boundaries marked by the black line, sufficient to identify those boundaries to a person approaching them, if at all material times, there were such marks, and provided that those marks are consistent with the black lines but not otherwise."

Mr. Howell, however, does not invite us to decide the case that way. On the contrary, he recognises that the validity of the Byelaws must be determined as at the date they are made and cannot depend upon the proper fulfilment of a duty which only arises thereafter - the section 17(1) duty to mark the boundaries on the ground. He accepts , therefore, that the court in Bugg was right to reject Mr. Bishop's argument - the argument that the Byelaws' failure to describe the boundary of the protected areas did not matter because of the section 17(1) duty. Mr. Howell's argument is rather that there was never in the first place any obligation to describe the boundaries in the Byelaws themselves. This, he points out, is an obligation expressly imposed by Parliament under section 2(5) of the 1900 Act with regard to particular (watery) areas, and for obvious good reason: it would not in those circumstances be right for the court to impose a similar requirement in all cases. Rather it is the area to which the Byelaws apply which has to be described, not the boundaries as such. That, of course, requires that the boundaries are capable of ascertainment but, submits Mr. Howell, this was sufficiently achieved here by the identification of the areas themselves in Byelaw 1. He so submits, indeed, even had no plan been annexed. Here, however, for good measure one was, and, although it was provided "for convenience of identification", it could nonetheless still properly be used to identify the limits of the protected areas -see Wiggington Ltd v Winster [1978] 1 WLR 1462, Graham v Philcox [1984] 1 QB 747, and Scott v Martin [1987] 1 WLR 841.

Mr. Pleming accepts that the boundaries themselves do not have to be described as such and further accepts that the plan is properly available to assist in identifying the protected areas. He disputes, however, the sufficiency of the description in Byelaw 1 with or without the plan. The limits of the protected areas must, he argues, be clear and certain and they must be readily ascertainable not merely by the Secretary of State but by any interested citizen. The boundaries, he concedes, need not necessarily be ascertainable within the four corners of the Byelaws themselves; if not, however, they must be ascertainable by reference to some available Ordnance Survey plan or central register or the like, to be notified by the Byelaws. At the very least, he argues, there should be more precise identification of the protected areas than was afforded by the 1986 Byelaws - as perhaps by the RAF Menwith Hill Byelaws 1996 which revoked the earlier Byelaws and which identify the new applicable area by reference to all the land shown on a different (and undoubtedly clearer) map "which lies within the outermost edges of the red lines on that map" - although even this is not conceded by Mr. Pleming to be sufficiently certain.

One of Mr. Pleming's points is that Byelaw 1 does not state even that it applies to all the lands belonging to the Secretary of State in the specified parishes. That clearly was not a point which troubled the court in Bugg - the judgment there referring simply to "the" lands belonging to the Secretary of State. Nor does it trouble me. Whatever else was achieved by the map, it at least served to depict the essential areas covered. Whether or not these were indeed all "the lands belonging to the Secretary of State" falling within the extended statutory definition of these terms - which Mr. Howell contends to be the natural meaning of Byelaw 1 - really matters not.

Mr. Pleming further argues that, because of the extended statutory definitions, there exist no means outside the Byelaws of checking just what lands do belong to the Secretary of State so as to be capable of being properly brought within the Byelaws. That certainly is true. The Secretary of State himself will of course know, but anyone else will have to ask him (and if necessary litigate) for the relevant information. That is equally true, however, with regard to the question whether the land has been properly appropriated for military use. I cannot see why the citizen's inability to discover either of these matters for himself unaided should be thought to create any particular uncertainty as to the limits of the areas designated as protected.

As to the plaintiffs' central complaint about uncertainty, upheld as this was in Bugg, namely that the precise boundaries of the protected areas are unclear either from their description or from the map, Mr. Howell points out that all maps, however good, ultimately leave an element of uncertainty: the lines on the map always have to be translated into lines on the ground. In reality the thickness of the line on the map and the scale of the map are likely to be of less importance than the lie of the land itself. In all cases it will be necessary to go to the land to find out what is represented by the map. Sometimes this will be obvious; other times, however apparently precise the map, less so.

If, submits Mr. Howell, there is indeed any uncertainty on the ground as to whether or not the Byelaws apply at a particular point, then the benefit of that doubt ought clearly be given to the citizen and he or she ought not to be convicted of a byelaw offence. That, however, is no reason to strike down the Byelaws in their entirety.

As to the obligation under section 17(1) to mark the boundaries of protected areas on the ground, this, Mr. Howell argues, is an important part of the statutory context within which to construe section 14(1) and determine the degree of descriptive certainty it requires. Again, if there is any failure by the Secretary of State to comply with his section 17(1) duty, or if the boundaries marked on the ground do not properly coincide with those of the area described in the Byelaws themselves, then in those respects the Byelaws will not be enforceable. That, indeed, is precisely what the Divisional Court held in Bugg with regard to the Alconbury Byelaws - see page 504C-D.

Those, then, are the rival contentions on the facts with regard to the question of certainty. What should be the legal approach?

Two competing approaches are canvassed, approaches which for convenience may be called respectively the Kruse v Johnson approach - see Kruse v Johnson [1898] 2 QB 91 - and the Fawcett approach - see Fawcett Properties Ltd v Buckingham CC [1964] AC 636.

The Kruse v Johnson approach for which Mr. Pleming contends is based on Mathew, J.'s dictum in that case:

". ....a byelaw to be valid must, among other conditions, have two properties - it must be certain, that is, it must contain adequate information as to the duties of those who are to obey, and it must be reasonable."

A more modern statement of essentially the same approach appears in Lord Lane, C.J.'s judgment in Staden v Tarjani (1980) 78 LGR 615:

".......to be valid, a byelaw carrying as this one does penalties for infringement, must be certain and clear in the sense that anyone engaged upon the otherwise lawful pursuit of hang gliding activity must know with reasonable certainty when he is breaking the law or is not breaking the law. That proposition scarcely needs demonstration or authority."

The rival approach urged by Mr. Howell is to be found in Lord Denning's speech in Fawcett:

"I can well understand a that a byelaw will be held void for uncertainty if it can be given no meaning or no sensible or ascertainable meaning. But if the uncertainty stems only from the fact that the words of the byelaw are ambiguous, it is well settled that it must, if possible, be given such a meaning as to make it reasonable and valid, rather than unreasonable and invalid."

Then this:

".....I am of opinion that a planning condition is only void for uncertainty if it can be given no meaning or no sensible or ascertainable meaning, and not merely because it is ambiguous or leads to absurd results. It is the daily task of the courts to resolve ambiguities of language and to choose between them; and to construe words so as to avoid absurdities or to put up with them."

A little later in the speech, in the course of indicating his preferred formulation as to the court's proper approach to reasonableness, Lord Denning said this:

".........it puts planning conditions on much the same footing as byelaws made by a local authority, to which they are so closely akin. Indeed, I see no difference in principle between them. As with byelaws so with planning conditions."

It appears from the beginning of the passage already cited from Bugg that the court there adopted the Kruse v Johnson approach; there had, indeed, been extensive citation of that line of authority whereas Fawcett and the subsequent planning condition cases had not apparently been before the court.

Mr. Howell submits that whichever approach the court adopts these particular Byelaws ought not properly to be condemned for uncertainty. Least of all, he submits, should they be struck down on the Fawcett approach. Mr. Pleming's submissions unsurprisingly lie at the opposite end of the spectrum. He argues that these Byelaws would fail even the less stringent test set by Fawcett: there is simply no sufficient information provided by them from which to ascertain the areas protected.

Early in the judgment though this is, I have to say that I find Mr. Howell's arguments compelling: even applying the Kruse v Johnson test I would uphold these Byelaws rather than condemn them for uncertainty. And this despite the considerable persuasive force inevitably attaching to the court's judgment in Bugg. Of course the plan here could have been better and clearer - as the 1996 plan now is. Of course too the boundaries could have been described - to include, for instance, just as in the original draft statutory instrument (but for whatever reason not in the Byelaws as made) "the outer perimeter walls and fences of HMS Forest Moor and Menwith Hill Station". But Parliament plainly cannot have envisaged generally such a description of the boundaries - see section 2(5); and however narrow and precise the line on a map, there will always be, literally, a borderline of uncertainty. That should not in my judgment invalidate the Byelaws and make them void and unenforceable even against those who deliberately and flagrantly trespass within the very centre of the protected areas.

Mr. Howell suggests that the Divisional Court's decision in Bugg - critical as it is of the Byelaws' failure to describe the boundaries of the protected areas - may have been given per incuriam of section 2(5), and certainly no reference to that provision appears anywhere in the report, not even in the argument. We are told, however, that it was referred to in counsel's skeleton arguments and I am therefore inclined to regard the decision as based on rather wider grounds. Alas, I find myself in respectful disagreement with it.

It follows from this central conclusion that neither of the two important and difficult points of law argued before us strictly falls for decision. Since, however, we were treated to full and able argument on each, and since it is not impossible that it may be sought to carry this litigation further, I for my part think it right to address them.

When byelaws are challenged for uncertainty, should the court follow the Kruse v Johnson approach or the Fawcett approach?

In support of the Kruse v Johnson approach Mr. Pleming points to nearly 100 years of authority in which Matthew, J.'s dictum has been cited with approval and apparently applied. It is, he acknowledges, a curiosity that so fundamental and entrenched a principle should find its origin in an obiter dictum in a dissenting judgment in a case concerned (like the three cases cited in support of the dictum) not with certainty but with reasonableness. Nevertheless, as early as 1901, in Nash v Findlay (1901) 18 TLR 92 Matthew, J.'s test was expressly applied by Lord Alverstone, C.J. sitting with two of the judges who had been in the majority in Kruse v Johnson, and a byelaw there was held void for uncertainty. Amongst the subsequent cases in which Matthew, J.'s dictum or an equivalent re-statement of it are to be found are Scott v Pilliner [1904] 2 KB 855, Leyton UDC v Chew [1907] 2 KB 283, Attorney General v Denby [1925] 1 Ch 596, United Bill Posting Co. Ltd v Somerset County Council (1926) 42 TLR 537, Staden v Tarjanyi (1980) 78 LGR 615 and R v Secretary of State for Trade & Industry ex parte Ford (1984) Trade LR Vol 4, 150, all of which we examined.

As for Fawcett, Mr. Pleming submits, first, that Lord Denning's dictum there, insofar as it equates byelaws with planning conditions, was itself obiter; and, second, that in any event Fawcett was concerned with uncertainty of language whereas here we are concerned with uncertainty of application. It is one thing, he submits, to require the courts to resolve the former; quite another for the courts to determine a byelaw's area of operation when the instrument itself lacks the necessary information. Lord Cohen's speech in Fawcett, he argues, supports only the former: explaining Viscount Simonds' dictum in London & North East Railway Co. v Berryman [1946] AC 278 that "a man is not to be put in peril upon an ambiguity", Lord Cohen said this:

"This principle involves that if a statutory provision is ambiguous, the court should adopt any reasonable interpretation which would avoid the penalty, but the court should not, I think, strike a provision out of an act on the ground of uncertainty unless it is impossible to resolve the ambiguity which it is said to contain."

Mr. Howell submits to the contrary that whether or not Lord Denning's dictum in Fawcett is obiter with regard to byelaws (which he disputes), planning conditions are closely analogous to them and no logical reasons exist for distinguishing between them. The breach of either can be attended by criminal consequences - admittedly, in the case of planning conditions, at one remove, but that is immaterial. If anything, submits Mr. Howell, one would expect a less stringent test of uncertainty for byelaws which must cater to a wide range of possibilities than for conditions which are directed to more specific circumstances.

As to Mr. Pleming's contention that the Fawcett principle in any event applies only to ambiguity of language rather than uncertainty of operation, Mr. Howell submits that this too would involve an illogical and unjustifiable distinction in approach. As Willmer,L.J. said in Hall & Co. Ltd v Shoreham UDC [1964] 1 WLR 240 at 245 - another planning condition case in which Fawcett was applied:

"It has been contended on behalf of the plaintiffs that conditions imposed by a local authority must be such that the developer can know exactly what they involve before making up his mind whether to proceed with the development contemplated.....I do not think, however, that this is the right test to apply. The question is not whether the plaintiffs are left uncertain as to how the conditions will operate. That may be very relevant in considering whether the conditions are reasonable. But where uncertainty is alleged, the true question, as it seems to me, is whether the language of the conditions make sense, that is, is capable of a reasonable construction. This is in accordance with the view expressed by Lord Denning in Fawcett...."

Similarly, as Woolf, J. observed in ex parte Ford, another statutory instrument case:

"As was made clear by a series of authorities, uncertainty of language rarely creates the necessary degree of invalidity to cause the courts to intervene. It seems to me that a similar approach has to be applied to uncertainty of application, and especially in the area of consumer protection it is unfortunately inevitable that those who are responsible for supplying the public will at times be put in the situation where they will have to make difficult decisions as to whether or not they are infringing the law."

If, clearly, the Fawcett principle applies equally to byelaws as to planning conditions, and to uncertainty of application as well as to uncertainty of language, there can be no question as to which line of authority we must follow: the planning condition cases are all in the House of Lords or Court of Appeal; the byelaw cases all in the Divisional Court or before a single judge. Fawcett has been followed in other planning condition cases in the Court of Appeal - Alderson v Secretary of State for the Environment (1984) 49 P&CR 307 - a case where the Court of Appeal overturned Webster, J.'s ruling that "locally" had no ascertainable meaning - as well as Hall v Shoreham UDC. Since Fawcett, however, Kruse v Johnson appears still to have been followed in other byelaw (or similar) cases -notably, of those we were shown, Staden, ex parte Ford, and, indeed, Bugg itself. Perhaps rather oddly, the inter-relation between these two strains of authority seems scarcely to have been considered: the only specific reference to it appears in Diplock, L.J.'s judgment in Mixnam's Properties Ltd v Chertsey UDC [1964] 1 QB 214, a case concerned with the conditions of a caravan site licence. At page 235 Diplock, L.J. said this:

"Failure to comply with a condition attached to a site licence is a criminal offence. The power to impose conditions is thus, in effect, a power to make subordinate legislation analogous to a power to make byelaws - see per Lord Denning in Fawcett. The validity of such conditions is thus to be tested by the same principle as the validity of byelaws."

Then, at page 238, this:

"Byelaws have in the past been declared void for 'uncertainty': see Nash v Findlay and Attorney General v Denby. Some doubt is cast on the correctness of 'uncertainty' as a separate ground of invalidity by the speeches in the House of Lords in the recent case of Fawcett; but if the courts can declare subordinate legislation to be invalid for 'uncertainty' as distinct from unenforceable, as in the case of a clause in a statute to which it is impossible to ascribe a meaning, this must be because Parliament is to be presumed not to have intended to authorise the subordinate legislation authority to make changes in the existing law which are uncertain."

The only conditions in fact struck down in Mixnam's as void for uncertainty were two which Danckwerts, L.J. found "quite impossible to apply", one of which Diplock, L.J. described as "so uncertain it is not possible to ascribe any intelligible meaning to the words"; the other he found ultra vires on a more general ground so that it was unnecessary to decide whether, because it was "difficult to ascribe [to it] any intelligible meaning", it was also void for uncertainty. Although I do not think that Mixnam's (which went on a different point to the House of Lords) decides the present issue, it seems to me broadly to support Mr. Howell's argument.

More importantly, however, Mr. Howell makes, to my mind convincingly, these two further submissions. First, that the Kruse v Johnson test - whether formulated in terms of "adequate information" (as by Matthew, J.), or of "reasonable certainty" (as in Staden) - itself is one of great uncertainty. It provides no criteria or principles by which to judge the adequacy of the information, or the degree of certainty, afforded by the byelaws. At what point does a byelaw become invalid for uncertainty? It seems to me insufficient to answer, as Mr. Pleming does, that this can safely be left to the magistrates of the day or the judges on appeal. Better, surely, as with the Fawcett test, to treat the instrument as valid unless so uncertain in its language as to have no ascertainable meaning, or so unclear in its effect as to be incapable of certain application in any case (often on analysis essentially the same thing - see for example Alderson). Being a criminal provision, it will be wherever reasonably possible construed - or, as the case may be, applied -to avoid penalty.

Mr. Howell's second and, as it seems to me, inter-related submission, is that, despite the language used and the approach ostensibly applied in all the byelaw cases, these, almost without exception, can be seen on examination to have produced a result entirely consistent with the Fawcett test. Some appear in the end not to be uncertainty cases at all - see, for example, Scott v Pilliner where, as Lord Denning pointed out in Fawcett, the byelaw was held void for unreasonabless, not uncertainty. Pressed to indicate which of the byelaw cases he contended would have been decided differently on a Fawcett approach, Mr. Pleming nominated Nash v Findlay, Staden, and Attorney General v Denby. In none of these cases, however, does that seem to me at all clear. In Nash v Findlay the court simply felt unable to ascribe any meaning whatever to a particular provision, given the existence of other byelaws which appeared to include the same area of misconduct. Staden was a case where the byelaw impugned forbade any person inter alia to hang-glide over a particular pleasure ground. Given that Bernstein v Skyviews [1978] QB 479 had just held it permissible to fly at such a height that no one can possibly be inconvenienced, Lord Lane, C.J. said this:

"It seems to me that to be valid the byelaw must set some lower level below which the glider must not fly. It is not for us to say whether that should be stipulated in feet or metres or whether it should be delineated by, if you like, the nuisance [or annoyance] to those on the ground......Some limitation must be put upon it....There was no such limitation. Accordingly, it follows that the byelaw was invalid."

The decision is explicable in terms of there having been no lawful touchstone of liability whatever. Had a height been set, that would have cured the vice of forbidding what was otherwise undoubtedly lawful flying. Had the criterion of nuisance and annoyance been used, uncertain in its application though that no doubt would have been , the court would still have found it acceptable.

Attorney General v Denby is a difficult case which I do not propose to discuss at length. Suffice it to say that the byelaw there in question was not in the event held invalid; rather, it was held "too uncertain for me to give effect to it in this case [my emphasis]; that on one construction there is no infringement, and on the other it is wholly unreasonable".

Certain other of the byelaw cases, moreover, strongly support the view that in reality the court, however it chose to express itself, was upholding byelaws even where they were undoubtedly of uncertain application. That seems to me conspicuously to have been the case in United Bill Posting v Somerset CC, and also in ex parte Ford. In United Bill Posting a byelaw was upheld which forbade the exhibiting of advertisements "so as to be visible from any public highway.....and to disfigure the natural beauty of the landscape". The empowering statute had referred to "the natural beauty of a landscape". The complaint that the byelaw did not specify the landscapes of natural beauty to which it applied (as clearly it could have done) failed.

Ex parte Ford was concerned with an order banning the sale of scented erasers, "scented" meaning "smelling of food or flowers". Woolf, L.J. noted that the application of the order gave rise to difficulties:

"The question as to whether or not an eraser.....does in fact smell of food or flowers must be very much a matter of subjective judgment. ....It is possible readily to envisage situations where different courts would come to different conclusions about the same eraser. That such a situation can be brought about by an order creating a criminal offence I have no hesitation in saying is highly undesirable."

The order was nevertheless upheld.

Why I describe this second submission of Mr. Howell's as inter-related to his first is this: it is surely for the very reason that a literal application of the Kruse v Johnson test would of itself involve great uncertainty that the courts in fact give effect to byelaws even when it is often quite impossible to be sure at the margins whether or not a particular course of conduct would breach them. In my judgment, so long as in certain circumstances an offence will undoubtedly be committed, byelaws should (subject to any issues of severance such as arose in Hutchinson) be upheld and to that extent enforced. If, of course, that test is failed, then indeed the byelaws should be struck down for uncertainty: Parliament could not have intended byelaws to be made which are incapable of any certain application. (There may, I recognise, be cases where byelaws are enacted which quite needlessly give rise to difficulties of application. Whether these should properly be struck down for unreasonableness it is not presently necessary to decide.)

From all this it follows that the Fawcett test of uncertainty should in my judgment apply to byelaws just as to planning conditions. There is no sufficient reason to distinguish between the two classes of case. The Fawcett test, moreover, is the more satisfactory of the two tests because itself the more certain. Even if not strictly bound to do so, therefore, I would adopt it as this court is free to do. On this approach - even assuming (which I respectfully doubt) that in reality it would generally make any practical difference to the result - there can be no doubt whatever that the present Byelaws fall to be upheld.

I come finally to the issue of lawful justification. If, contrary to my own clear conclusions, Bugg was right to hold these Byelaws invalid for uncertainty, can they nevertheless provide a defence of lawful justification to the tortious claims now faced by these defendant constables for the arrests and detentions which they carried out in purported enforcement of them?

Mr. Pleming submits not, taking as his starting point Lord Lowry's speech in DPP v Hutchinson:

"The basic principle is that an ultra vires enactment, such as a byelaw, is void ab initio and of no effect."

That is, indeed, hardly a controversial proposition. Another authoritative statement to the same effect appears in Lord Diplock's speech in Hoffmann-La Roche v Secretary of State for Trade and Industry [1975] AC 295 at 365 (a case, as we shall see, upon which Mr. Howell nevertheless places considerable reliance):

"It would....be inconsistent with the doctrine of ultra vires as it has been developed in English law as a means of controlling abuse of power by the executive arm of government if the judgment of a court in proceedings properly constituted that a statutory instrument was ultra vires were to have any lesser consequence in law than to render the instrument incapable of ever having had any legal effect upon the rights or duties of the parties to the proceedings....."

Invalidity is, in short, generally regarded as of retrospective not prospective effect.

Given that section 17(2) of the 1892 Act (and Byelaw 4) permit arrest only if the person arrested actually commits an offence, it would, submits Mr.Pleming, be fundamentally inconsistent with this basic principle now to regard the defendant constables as having acted lawfully.

To that apparently formidable argument Mr Howell makes two main responses, either, he submits, capable on its own of defeating it; taken together, more compelling still.

The first centres on the decision of the House of Lords in Wills v Bowley [1983] AC 57. That case concerned a constable's power of arrest under section 28 of the Town Police Clauses Act 1847, a section which required a constable to "take into custody, without warrant, and forthwith convey before a justice, any person who was in his view [meaning "in his sight"] commits any such offence [a large spectrum of street offences, some serious some trivial, then being set out]".

By a majority of three to two the House of Lords held that section 28 falls to be construed so as to protect constables who arrest someone whom they honestly (albeit mistakenly) believe on reasonable grounds they have seen committing a stipulated offence. Lord Bridge said this:

"Parliament, in enacting any such provision, must have intended that any person who was committing any of the specified offences, whether serious or trivial, should be arrested and brought to justice, very often, no doubt, because this might be the only way he could be brought to justice at all. But the person making the arrest cannot determine guilt in advance; he cannot know that guilt will in due course be established; his only protection, if he is to have any, at the time of making the arrest must be found in his honest belief on reasonable grounds that he has observed the commission of a relevant offence by the person he arrests. If a power of arrest in flagrante delicto is to be effective at all, the person who exercises it needs protection; protection only against liability to pay damages in tort, but, perhaps more important, as the instant case shows, protection, so far as the law can give it, against violent resistance to the reasonable force which a person exercising a lawful power of arrest is entitled to use in order to effect and maintain his arrest. If the protection the law affords is contingent and unpredictable, how can Parliament reasonably have expected anyone to rely on it? Yet, surely Parliament must have intended the protection to be relied on in order that the power of arrest should be effective. Making an arrest can never be an agreeable task and may often be very disagreeable; how much more so if the law gives no assurance of protection."

By the same token, submits Mr. Howell, the constables here should be protected.

Not so, submits Mr. Pleming, for what I understand to be two main reasons: first, because of certain differences between the requirements respectively of section 28 there and of section 17(2) here; second, because Wills v Bowley is concerned only with honest and reasonable mistakes of fact, not of law. Let me consider each in turn.

The main difference in the legislation on which Mr. Pleming relies is that whereas under section 28 of the Town Police Clauses Act the constable (who, incidentally, had no option but to arrest the miscreant) was bound without more ado then to bring him before a court, here under section 17(2) the constable has a discretion not merely whether or not to arrest someone apparently in breach of the byelaws, but also, even if he does arrest him, whether then to bring him before a court or instead release him without charge; in practice, we are told, many are arrested and removed and not then prosecuted. True, Mr. Pleming acknowledges, here as in Wills v Bowley the legislation encompasses a wide range of offences, the most serious of which no doubt involve considerations of public safety and security. At the other end of the spectrum, however, are offences as venial as affixing leaflets to the outer side of the perimeter walls and fences. Where such a spectrum of seriousness exists, he submits, Lord Bridge's speech, properly understood, requires that before adopting the benevolent construction adopted there, the legislation must as there require that any offender will be automatically taken before a justice.

I would reject this argument. There seems to me in this case, no less than in Wills v Bowley, an obvious need for prompt action to secure public safety and security with regard to land appropriated for military use, and the mere fact that a constable is not bound to bring everyone arrested before the court ought not in principle to distinguish the two cases: section 17(2) involves essentially the same three sequential steps as section 28. In the result, had a constable arrested and ejected someone whom he genuinely (although on the facts mistakenly) believed on reasonable grounds to be a trespasser, I see no good reason why a Wills v Bowley construction of section 17(2) should not be adopted.

What then of the argument that such a construction cannot avail an arresting officer whose mistake was, as postulated here, one of law rather than fact?

Clear it is that ordinarily a mistake of law, however understandable, cannot provide the lawful justification for an arrest where otherwise there is none - see, for example, Bentley v Brudzinski (1982) 75 CAR 217 and Collins v Wilcock [1984] 3 AER 374 and Todd v DPP [1996] Crim LR 344. Here, however, submits Mr. Howell, looking at the matter as at the dates of these arrests, there was no mistake of law on the part of the arresting constables, certainly not in any conventional sense. It was not as if the constables had, as in the usual case, misunderstood their legal powers; on the contrary, they were enforcing what at the time appeared to be perfectly valid byelaws; to have done otherwise would seemingly have involved them in a clear breach of their duties. This essentially is Mr. Howell's second main argument and in support of it he relies heavily on Hoffmann-La Roche.

The question arising there was whether the Crown should be required to give a cross undertaking in damages when seeking an interlocutory injunction to enforce a statutory order which was under challenge on natural justice grounds. In holding not, it was emphasised by several of their Lordships that there is a duty to enforce the law and that the order had the full force of law unless and until it could be shown to be ultra vires. As Lord Reid put it:

"Dealing with alleged breaches of the law is a function of the Crown (or of a department of the executive) entirely different in character from its function in protecting its proprietary right. It has more resemblance to the function of prosecuting those who are alleged to have committed an offence. A person who is prosecuted and found not guilty may have suffered serious loss by reason of the prosecution but in general he has no legal claim against the prosecutor. In the absence of special circumstances I see no reason why the Crown in seeking to enforce orders of this kind should have to incur legal liability to the person alleged to be in breach of the order. It must be borne in mind that an order made under statutory authority is as much the law of the land as an act of Parliament unless and until it has been found ultra vires."

Lord Diplock at page 367, said this:

"The duty of the Crown to see that the law declared by the statutory instrument is obeyed is not suspended by the commencement of proceedings in which the validity of the instrument is challenged. Prima facie the Crown is entitled as of right to an interim injunction to enforce obedience to it. To displace this right or to fetter it by the imposition of conditions it is for the defendant to show a strong prima facie case that the statutory instrument is ultra vires."

Mr. Pleming's response to this argument is twofold. First, he reiterates the basic principle as to the retrospective effect of any eventual finding of invalidity. Second, he submits that in this particular case the Byelaws were patently defective, i.e. invalid on their face.

Mr. Pleming seeks to illustrate the operation of the basic principle in this field by reference to such authorities as R v Reading Justices ex parte South West Meats Ltd (1992) Crim LR 672 and R v Central Criminal Court ex parte AJD Holdings Ltd (1992) Crim LR 669. In my judgment, however, neither avail him: critical to each was the provision in section 15(1) of PACE that "an entry on or search of premises under a warrant is unlawful unless it complies with this section and section 16 below". Once the court had found on the facts in each case that they did not so comply, the entry and search were necessarily held unlawful. The cases simply did not concern the question whether there is legal justification for those things done by a person discharging his duty to enforce the apparent law if subsequently that law proves to be invalid. Nor to my mind does Mr. Pleming's argument derive any assistance from consideration of the statutory protection given to justices by the Justices of the Peace Act 1979 as amended, or to constables by the Constables Protection Act 1750.

Mr. Pleming's second argument, that the Byelaws here were patently defective, it is necessary to consider in a little more detail. Once again it takes as its starting point the decision in Bugg. The Divisional Court there, Mr. Pleming points out, speaks of the present class of case (which it held susceptible of collateral challenge) as involving "substantive invalidity" (as opposed to "procedural invalidity"), where the byelaw is "on its face invalid" (ie where "no evidence is required", as for example, where it is "patently unreasonable"). And the court at page 500 added this:

"A member of the public is required to comply with byelaws even if he believes they have a procedural defect unless and until the law is held to be invalid by a court of competent jurisdiction. If before this happens he contravenes the byelaw, he commits an offence and can be punished. Where the byelaw is substantively invalid, the position is different. No citizen is required to comply with a law which is bad on its face. If the citizen is satisfied that that is the situation, he is entitled to ignore the law."

This approach, moreover, was followed and applied in R v Wickes (1995) 93 LGR 377, an enforcement notice case where the CACD, returning to the language of Miller-Mead v MHLG [1963] 2 QB 196, described Woolf, L.J.'s "substantive invalidity" as akin to "nullity", and his "procedural invalidity" as relating to such matters (which "would in almost all cases require ..... evidence") as would render a notice invalid. (The actual decision in Wicks was that the notice there, not being a nullity patently defective on its face, remained valid until quashed, something which only the High Court rather than the criminal court had jurisdiction to do.)

This argument too, however, I would reject. By no stretch of the imagination can these Byelaws be regarded as so patently defective on their face that the constables can properly be criticised for not having identified their deficiencies and declined to enforce them. The instrument, in Lord Radcliffe's celebrated phrase in Smith v East Elloe RDC [1956] AC 736 at 769, "bears no brand of invalidity upon its forehead."

The same Divisional Court as eventually in Bugg held these byelaws invalid for uncertainty had, after all, the previous year in Greaves regarded that very contention as unarguable.

I can quite understand the concepts of "substantive invalidity", "procedural invalidity" and so forth in the context of determining whether or not collateral challenges properly lie. In my judgment, however, they afford no help whatever in the present situation. There can be no question here of the constables lacking "reasonable grounds" for believing these plaintiffs to have been breaching the law (the Wills v Bowley approach), nor of there being apparent to them "a strong prima facie case that the statutory instrument [was] ultra vires" (Lord Diplock's phrase in Hoffmann-La Roche). As it seems to me, the position can be no different whether byelaws are eventually condemned for substantive invalidity or for procedural invalidity: either way, as I repeat, the actions of the constables concerned - who might theoretically have heard the court dismiss Miss Greaves' judicial review application as unarguable - would be beyond criticism. (Having reached this conclusion I note but need not deal with Mr. Howell's submission that, even by Bugg's own test of patent invalidity (i.e. where no evidence is required to establish it), absent the concession there made as to the width of the line on the plan representing 60 feet on the ground, evidence would in any event have been required so that the test is not satisfied here.)

Although I do not pretend to have found this part of the case altogether easy, I have come to the conclusion that Mr. Howell's arguments are to be preferred here also.

The central question raised here is whether these constables were acting tortiously in arresting the plaintiffs or whether instead they enjoy at common law a defence of lawful justification. This question, as it seems to me, falls to be answered as at the time of the events complained of. At that time these Byelaws were apparently valid; they were in law to be presumed valid; in the public interest moreover, they needed to be enforced. It seems to me one thing to accept, as readily I do, that a subsequent declaration as to their invalidity operates retrospectively to entitle a person convicted of their breach to have that conviction set aside; quite another to hold that it transforms what, judged at the time, was to be regarded as the lawful discharge of the constables' duty into what must later be found actionably tortious conduct.

I do not understand this point ever to have been addressed before. In my judgment it is not covered by the general doctrine of retrospectivity with regard to the annulment of invalid instruments. I am not prepared to regard the many broad statements of principle as going this far.

On the face of it, any right of redress on the part of those arrested under what ultimately are found to be defective byelaws should be against the Secretary of State as the maker of the invalid instrument. The Secretary of State is, indeed, here said to be liable on that ground. If, however, as Mr. Pleming recognises may well be the case, no such claim succeeds, essentially because English law provides no cause of action for invalid administrative action as such, that is no basis for creating a cause of action instead against those, here the defendant constables, who are not responsible for the invalidity. Nor is it a good ground for denying them the common law defence of lawful justification which should surely be available to them. Quite the contrary. I see no sound policy reasons for making innocent constables liable in law, even though such liability would be underwritten by public funds.

In my judgment, therefore, even if these Byelaws are properly to be regarded as void for uncertainty as the Divisional Court held in Bugg, that would not serve to deprive the constables here of a defence of lawful justification wherever they can show they were acting in the reasonable belief that the plaintiffs were committing a byelaw offence.

Rather than reformulate the preliminary issues yet again, I would suggest that this be treated as a declaratory judgment.

LORD JUSTICE PETER GIBSON: The preliminary issues determined by the Judge were essentially : (1) Are the byelaws defective on their face on the ground that the area to which they relate is insufficiently identified, and, if so, are the byelaws invalid? (2) If the byelaws are invalid, were they incapable of providing a lawful justification for the arrest and detention of the Plaintiffs? Both issues raised questions of general importance.

Byelaws are a widely used form of delegated legislation, a usual feature of which is that they prescribe what may or may not be done within a designated area and attach a penal sanction to any breach. They are to be found regulating, for example, the parking in streets in a local authority's area and regulating commons, parks and communal gardens. If therefore the plaintiffs are right in their contention that the byelaws with which this case is concerned are uncertain and therefore invalid because they fail to provide adequate information as to the protected area, the principle thereby established would have wide application.

Mr. Pleming Q.C. in his powerful argument for the Plaintiffs naturally placed reliance on the judgment of the Divisional Court in Bugg v D.P.P. [1993] Q.B. 473, to the effect that the byelaws were defective on their face because there was no description contained in the byelaws as to the boundary of the protected area. If the phrase "lands belonging to the Secretary of State" in the specified parishes means the lands belonging to the Secretary of State, as I believe, and the Divisional Court appears to have believed, it does, then in my judgment the protected area has been described with certainty. It is not suggested that the Secretary of State is unable to establish that the lands shown on the plan do belong to him. And if the area is certain, it follows that the boundary of the area is certain. Further, the reference to the plan for convenience of identification means that the plan, whilst it cannot override the verbal description of the lands as belonging to the Secretary of State in the specified parishes, can be used as an aid to determining the protected area and hence its boundary. In my judgment, this not being an area to which s.2(5) Military Lands Act 1900 applied, it was not necessary to describe the boundary in the byelaws. I therefore respectfully disagree with the Divisional Court in so far as their decision is based on the absence of a description of the boundary of the protected area in the byelaws.

Mr. Pleming in arguing for the invalidity of the byelaws submitted that the description "lands belonging to the Secretary of State" made it impossible for anyone walking in the area to know what land is referred to. I cannot accept that that is the right test. Mr. Pleming was constrained to accept that the protected land might properly be described in the byelaws by reference to an Ordnance Survey plan or public register, but he submitted that the extrinsic evidence must be readily available to the public. I cannot see how this can properly be an ingredient of a test of certainty. So long as the description identifies the protected area by whatever means, the test of certainty is in my judgment satisfied. No doubt one of the reasons why Parliament required in s.17 Military Lands Act 1892 the marking of the boundaries of the areas to which byelaws apply was that members of the public in the neighbourhood of the protected area should know where the area began, but that marking, as the Divisional Court in Bugg v D.P.P. rightly held, did not affect the question of certainty.

Another point taken by Mr. Pleming is that uncertainty may be caused by the enlargement or reduction, after the byelaws are made, of the lands belonging to the Secretary of State. For my part, I do not accept that such enlargement or reduction creates any uncertainty. The meaning of the byelaws is that which they have when the byelaws are made. If the Secretary of State acquires more land, the byelaws will not apply to the further land; that would require a further decision by the Secretary of State to be taken in the light of the circumstances then obtaining to make byelaws apply to the further land. If the Secretary of State disposes of land subject to the byelaws, the land disposed of would no longer be subject to the byelaws, as they would cease to satisfy the condition of belonging to the Secretary of State. Further the plan would need amending, and no doubt in practice there would be new byelaws promulgated for the smaller area.

In my judgment, if the correct test for the certainty of a byelaw is, as Mr. Pleming submitted, that which was suggested by Mathew J. in Kruse v Johnson [1898] 2 QB 91 at p.108, viz. that it must contain adequate information as to the duties of those who are to obey, then in my judgment these byelaws passed that test. For these as well as the reasons given by Simon Brown L.J. I would allow this appeal.

On the further point whether the correct test is that suggested by Mathew J. or that suggested by Lord Denning in Fawcett Properties Ltd. v Buckingham County Council [1964] A.C. 636 at pp. 676-8 (the most material parts of which Simon Brown L.J. has cited), I unhesitatingly prefer that of Lord Denning. The adequacy of information test is itself uncertain. Our attention has been drawn to Bennion : Statutory Interpretation 2nd ed. (1992) 326 where it is stated :

"Voidness for uncertainty The interpreter is not permitted to declare an enactment containing a broad term or other ambiguous expression void for uncertainty. The uncertainty is intended to be resolved by the interpreter, whether an administrative official or the court. This can be looked on as a delegation by Parliament of legislative power. A corresponding rule applies to uncertainty in statutory instruments and most other delegated legislation.

The position is different with byelaws however. A byelaw may be declared void if uncertain in its terms."

To my mind it is wrong in principle that the test of uncertainty should differ according to whether it applies to a byelaw on the one hand or to an enactment or delegated legislation other than a byelaw on the other. Further it should be borne in mind that in Kruse v Johnson itself, Lord Russell C.J. (with whom Chitty L.J. and Wright, Darling and Channell JJ. agreed) urged (at p.99) that byelaws made by representative public bodies ought to be supported, if possible, and interpreted benevolently. There is no dispute but that if no meaning can be given to a byelaw, it must be invalid, and that if there is any reasonable doubt whether an unauthorised person has entered the protected area, that person will not be convicted of an offence under the byelaws. But where a sensible meaning can be given to the byelaws, and there is no reasonable doubt that the unauthorised person is within the protected area, I can see no sufficient reasons of policy or principle that should lead the court to conclude that the byelaws should be struck down on the ground of uncertainty.

For these reasons and the reasons explained by Simon Brown L.J. in his judgment I too would hold that the Fawcett test for uncertainty is to be preferred to that suggested by Mathew J. in relation to byelaws.

On the second issue, I am in entire agreement with Simon Brown L.J.

LORD JUSTICE SCHIEMANN: I agree with both judgments which have been delivered. The use of the phrase "lands belonging to the Secretary of State" in the Byelaws does not render them uncertain because whether or no lands do or do not belong to the Secretary of State can be ascertained with certainty - by resort to the courts if need be. The fact that it is conceivable that an individual walking in the area might well have difficulty in knowing, without making further inquiries, whether a particular square metre did or did not belong to the Secretary of State is, in my judgment irrelevant to the question whether or no the Byelaws are void for uncertainty.

I desire only to add a few words on what the position would have been had the court found the byelaws in question to be invalid.

Underlying this is a question which has intrigued administrative lawyers in different countries and to which they have given different answers. The question is this : once a court has declared an enactment to be invalid, from what point in time does the abrogation apply, retroactively from the time of the court's ruling (ex tunc) or only from the time of the court's ruling (ex nunc). Historically, this has been a problem of far greater import in countries which have a court which can declare legislative acts void as being unconstitutional than it has been in this country where traditionally it has only been the legality of bye laws and statutory instruments which has been the subject of legal challenge. Now that we are members of the European Union and the possibility arises that even provisions in Acts of Parliament can be declared illegal because of a conflict with Comunity law the question may well grow to be of greater importance in this country.

The ex tunc solution has an initial attractiveness. The law should never have been made and therefore one must proceed as though it never had been made. To do otherwise will in effect legalise the illegal and the courts are not in business to do that. Moreover, once the courts start to give some effect to illegal legislation, there will be less incentive for the legislator to refrain from such illegality.

The problem with the ex tunc solution is that it will often be the case that, between the making of the enactment in question and the declaration of its invalidity, many people will have regulated their lives on the assumption that the enactment was lawful. Society cannot function if all legislation has first to be tested in court for legality. In practice, money will have been spent, taxes collected, businesses and property bought and sold and people arrested and perhaps imprisoned on the basis that what appears to be the law is the law.

It has been commonplace in our jurisprudence, as Simon Brown LJ points out, to speak of a basic principle that an ultra vires enactment is void ab initio and of no effect. This beguilingly simple formulation, as is widely acknowledged, conceals more than it reveals. Manifestly in daily life the enactment will have had an effect in the sense that people will have regulated their conduct in the light of it. Even in the law courts it will often be found to have had an effect because the courts will have given a remedy to a person disadvantaged by the application of the ultra vires enactment to him or because a decision, binding on the parties thereto, has been rendered on the basis of the apparent law or because some period of limitation has expired making it too late now to raise any point on illegality.

The policy questions which the law must address in this type of case is : whether any and if so what remedy should be given to whom against whom in cases where persons have acted in reliance on what appears to be valid legislation. To approach these questions by rigidly applying to all circumstances a doctrine that the enactment which has been declared invalid was "incapable of ever having had any legal effect upon the rights and duties to the parties" seems to me, with all respect to the strong stream of authority in our law to that effect, needlessly to restrict the possible answers which policy might require. For instance, in the context of the present case had we held the Byelaws to be invalid, as it seems to me a sensible answer might be that no remedy should be given against the policeman who acted in good faith but that a remedy should be given against a Secretary of State who put into circulation invalid byelaws with the perfectly foreseeable consequence that policemen would act on the assumption that they were valid. We are not required in the present case to decide whether that is the answer given by our law as it stands. I merely give it as a possibility.

It may be that, in the development of the law, future cases will draw on that part of our law which is applicable to cases containing a European Community element which shows a considerable amount of flexibility in dealing with this question. There are now many cases which examine the conflict which an ex tunc declaration produces with the principles of legal certainty, acquired rights and legitimate expectation. This is not the place to develop the point, there having understandably and sensibly been no reference to it at the bar and it being unnecessary to our decision. Relevant case law and discussions of the problems involved can be found, in the European context, in Wyatt & Dashwood's "European Community Law"(3rd Ed.) pages 93-95, Schermers & Waelbroeck "Judicial Protection in the European Communities" (5th Ed.) pages 56-63,441,442, in the English context in PP Craig "Administrative Law" pages 472-475, and, in the French context, in "Les grands arrets de la jurisprudence administrative" (9th edition by M. Long et al)p565-567.

So far as the present case is concerned, I would allow the appeal.

Order: appeal allowed; legal aid taxation of costs; leave to appeal refused.