//
© 2026 Zoe. All rights reserved.
Zoe is a legal information platform. Always consult the Federal Register of Legislation for authoritative text.
[1996] HCA 48
High Court of Australia
1996-03-28
Brennan CJ, Gummow JJ
Original judgment source is linked above.
High Court of Australia Brennan CJ Dawson, Toohey, Gaudron, McHugh and Gummow JJ McGinty v Western Australia (FC 96/001) [1996] HCA 48
ORDER
(i) Is s 6 of the Constitution Acts Amendment Act 1899 WA invalid?
(ii) Are ss 2A(2), 6 and 9 of the Electoral Distribution Act 1947 WA, or any of them, invalid?
(iii) If any of ss 2A(2), 6 or 9 of the Electoral Distribution Act 1947 WA is or are invalid, is it or are they severable and, if so, to what extent?
The plaintiffs, Messrs McGinty and Gallop, are Members of the Legislative Assembly of Western Australia (the Assembly). The plaintiff Mr Halden is a Member of the Legislative Council of that State (the Council). They challenge the validity of the laws of Western Australia governing the distribution of electorates for both the Assembly and the Council. The present distribution was effected pursuant to the Constitution Acts Amendment Act 1899 WA (the 1899 Act) and the Electoral Districts Act 1947 WA (the 1947 Act) as amended by the Acts Amendment (Electoral Reform) Act 1987 WA (the 1987 Act). The 1947 Act was renamed the Electoral Distribution Act [1] . The Assembly now consists of fifty-seven members [2] each of whom is returned as a member for an electoral district. The fifty-seven electoral districts [3] are divided between the Metropolitan Area containing thirty-four electoral districts [4] and the remainder of the State containing twenty-three electoral districts [5] . The districts are established by Electoral Distribution Commissioners who are required to fix the boundaries of the districts in each area so as to comprise an equal quotient of enrolled electors plus or minus 15 per cent [6] . The Metropolitan Area is defined [7] as the region that was described at 1 January 1987 in the 3rd Sch to the Metropolitan Region Town Planning Scheme Act 1959 WA and Rottnest Island.
At the time when the 1987 Act was enacted, there were approximately 669,293 voters enrolled in the Metropolitan Area to elect thirty-four members and 240,081 voters enrolled in the remainder of the State to elect twenty-three members of the Assembly. The consequence of this apportionment was to confer on a voter in an average metropolitan electoral district power to effect the return of a member of the Assembly that was less, in mathematical terms, than the power conferred on a voter in an average non-metropolitan electoral district. The 15 per cent tolerance from the average quotient and shifts in population since the last distribution have produced some large divergences in the number of voters enrolled for districts in the two areas in the 1993 election. At the 1993 election, the quotient for a metropolitan electorate was 21,988 and Wanneroo, the most populous electorate, had 26,580 enrolled voters. The quotient at the 1993 election for an electorate outside the metropolitan area was 11,702 and Ashburton, the least populous electorate, had only 9,135 enrolled voters. The number of Wanneroo voters was 291 per cent of the number of Ashburton voters.
The Legislative Council consists of thirty-four members returned for six electoral regions [8] . The names of these regions and the number of members of the Legislative Council returned by each are as follows [9] : North Metropolitan Region 7 members; South Metropolitan Region 5 members; East Metropolitan Region 5 members; South West Region 7 members; Agricultural Region 5 members; Mining and Pastoral Region 5 members.
The Electoral Distribution Commissioners are required to divide up the State into these six regions within which there are to be complete and contiguous districts [10] . The first three of these regions together form the Metropolitan Area [11] . The Mining and Pastoral Region is remote from the capital where the land use is primarily for mining and pastoral purposes [12] . The Agricultural Region is generally south, or south and west, of and adjacent to the Mining and Pastoral Region and the South West Region consists of the remainder of the State [13] . At the 1993 election, the quotients applicable to the respective Council regions were as follows: North Metropolitan 34,161; South Metropolitan 33,876; East Metropolitan 32,822; South West 13,721; Agricultural 13,161; Mining and Pastoral 9,097. The number of voters in the quotient for the North Metropolitan Region was 376 per cent of the number of voters in the quotient for the Mining and Pastoral Region.
The disparity between the number of enrolled voters in the metropolitan and non-metropolitan districts for the election of members of the Assembly and the disparity between the number of enrolled voters in the several regions for the election of members of the Council provide the factual foundation for the challenge to the validity of the provisions of the 1987 Act which prescribe the current electoral distribution. Those disparities, it is said, are the consequence of the 1987 Act. To an extent, some disparities are mandated by the 1987 Act's division of the State into electoral districts and regions, and increases in the extent of those disparities are authorised by the tolerance on quotients authorised by the 1947 Act.
The plaintiffs submit that disparities in voting power are inconsistent with the principle of representative democracy as that principle is understood at the present time. Representative democracy, so the argument runs, requires that (a) every legally capable adult has the vote; and (b) each person's vote be equal to the vote of every other person. Of course, the term "legally capable adult" assumes without defining the scope of the franchise. In this century, the age of legal adulthood has been reduced from twenty-one to eighteen and the legal incapacity of women to vote has been removed. Aborigines, who were once constitutionally disqualified from the franchise, are no longer so disqualified. But age, sex and race are not the only qualifications that have governed an adult's right to vote. Other qualifications have related to ownership of property and education or a period of residence within the electoral district. Disqualifications still include the status of convicted criminal and mental infirmity or absence from registered address. In view of the fact that the franchise has historically expanded in scope, it is at least arguable that the qualifications of age, sex, race and property which limited the franchise in earlier times could not now be reimposed so as to deprive a citizen of the right to vote [14] .
The qualification of electors is one thing; the voting power of those who are qualified to vote is another. In the Commonwealth Constitution, for example, s 30 was for a time the provision determining the franchise for the election of the House of Representatives. But s 26 prescribed the number of members to be chosen by the respective States and s 29 empowered State Parliaments to make laws for determining electoral divisions within the State.
The term "representative democracy" implies that the franchise be so general in its scope and voting power be so distributed among those who have the franchise that those who are elected to govern can fairly be seen to be representatives of the people who are governed. Stephen J said in Attorney-General (Cth); Ex rel McKinlay v The Commonwealth [15] :
It is no doubt true that something approaching numerical equality of electors within electorates is an important factor, together with much else, in the attainment of what many will regard as representative democracy in its purest form, just as adult suffrage, free of discrimination on the grounds of race, sex, property or educational qualification will likewise aid in its attainment. But neither of these in absolute form is necessarily imported into the Constitution by the selection of representative democracy as the chosen mode of government for the nation. 1. (1975) 135 CLR 1 at 57.
The plaintiffs submit that both the Commonwealth Constitution and the Western Australian Constitution incorporate representative democracy as the central principle of government. They challenge the proposition, accepted in McKinlay [16] , that an equality of voting power among the electors possessing the franchise is not mandated by the Commonwealth Constitution. Western Australia, on the other hand, submits that neither the Commonwealth Constitution nor the Western Australian Constitution requires equality of voting power. Western Australia points out that an electoral system may give effect to the political will of the overall majority of voters although there be a disparity in voting power as between the voters in some districts or regions and the voters in other districts or regions, that representation of minority groups or adequate representation of geographical districts in the Parliament might be secured only, or more efficiently, by allowing a disparity of voting power [17] and that some disparity may assist in the formation of stable political institutions (both legislative and executive). These points may be well taken, although the argument focuses on the political results that can follow an unequal distribution of voting power rather than on an equality of voting power possessed by each holder of the franchise. The question whether differences in voting power can be justified by distinctions based on political opinion, minority interests or geographical residence does not admit of a definitive answer.
However, it is unnecessary and, for reasons presently to be stated, impermissible to determine the validity of the 1987 Act by reference to its consistency with the requirements of a general principle of representative democracy. "Representative democracy" has been used as a shorthand description of the form of government prescribed by the Commonwealth Constitution in order to explain how the freedom to discuss governments and political matters is implied in the Constitution. As "the people" are to choose their elected representatives [18] , it has been held that the people must be left free to discuss political and economic matters in order to perform their constitutional functions [19] .
Implications are not devised by the judiciary; they exist in the text and structure of the Constitution and are revealed or uncovered by judicial exegesis [20] . No implication can be drawn from the Constitution which is not based on the actual terms of the Constitution, or on its structure [21] . However, as an implication will be applied in a particular case to a specific factual situation, it may be expressed in terms relevant to that situation [22] . Although the Court was divided in Australian Capital Television Pty Ltd v The Commonwealth (ACTV) , there was nothing in any judgment to cast doubt on the approach then taken by Mason CJ [23] :
It may not be right to say that no implication will be made unless it is necessary. In cases where the implication is sought to be derived from the actual terms of the Constitution it may be sufficient that the relevant intention is manifested according to the accepted principles of interpretation. However, where the implication is structural rather than textual it is no doubt correct to say that the term sought to be implied must be logically or practically necessary for the preservation of the integrity of that structure. 1. See Nationwide News (1992) 177 CLR 1 at 41-45 and ACTV (1992) 177 CLR 106 at 133-136and the cases cited in those passages. 2. Amalgamated Society of Engineers' v Adelaide Steamship Co Ltd (the Engineers' Case) (1920) 28 CLR 129 at 145, 155; Melbourne Corporation v The Commonwealth (1947) 74 CLR 31 at 83 ACTV (1992) 177 CLR 106 at 135, 209-210. 3. ACTV (1992) 177 CLR 106 at 158-159 4. ACTV (1992) 177 CLR 106 at 135
In Nationwide News Pty Ltd v Wills (Nationwide News) [24] and ACTV the legislative power of the Parliament to enact the respective laws impugned in those cases was in question. In the former case, that question was answered by some Justices without reference to any implication affecting the scope of the legislative power invoked to support the law. But, relevantly for present purposes, an implication relating to freedom of communication, expressed in different terms in the several judgments, was held by some Justices in the former case and by a majority in the latter case to deny validity to the impugned laws. The implication, albeit differently expressed, was drawn from the text and structure of Pts II and III of Ch I of the Constitution and, in particular, from the provisions of ss 7 and 24. These provisions were seen to prescribe a manner of choosing members of the Senate and of the House of Representatives which would have been subverted if the impugned laws had validly imposed the restrictions on communications which they prescribed. In expounding the manner of choosing members of the Senate and House of Representatives and in showing how the impugned legislative restrictions would subvert that manner of choice, the term "representative democracy" or "representative government" was employed.
Although the term "representative democracy" is useful to explain the text on which the implied freedom depends, the term is not to be found in either the Commonwealth Constitution or the Western Australian Constitution. It is logically impermissible to treat "representative democracy" as though it were contained in the Constitution, to attribute to the term a meaning or content derived from sources extrinsic to the Constitution and then to invalidate a law for inconsistency with the meaning or content so attributed. The text of the Constitution can be illuminated by reference to representative democracy but the concept neither alters nor adds to the text. In Victoria v The Commonwealth [25] , Windeyer J identified the role of implications in constitutional interpretation:
Implications of The Moorcock kind give content and consequences to written contracts. Implications of a different kind may aid the interpretation of statutory provisions. That is because these must always be read as parts of a whole and with due regard to the subject with which the statute deals. In each case an implication means that something not expressed is to be understood. But in the one case this involves an addition to what is expressed: in the other it explains, perhaps limits, the effect of what is expressed. It is in the latter sense that, in my view of the matter, implications have a place in the interpretation of the Constitution: and I consider it is the sense that Dixon J intended when in Australian National Airways Pty Ltd v The Commonwealth [26] he said: "We should avoid pedantic and narrow constructions in dealing with an instrument of government and I do not see why we should be fearful about making implications." The only emendation that I would venture is that I would prefer not to say "making implications", because our avowed task is simply the revealing or uncovering of implications that are already there. The text, however illuminated by implication in one case, remains to govern decision in the next. 1. (1971) 122 CLR 353 at 401-402. 2. (1945) 71 CLR 29 at 85.
The principle of "representative democracy" can be given the status of a constitutional imperative [27] , but only in so far as the meaning and content of that principle are implied in the text and structure of the Constitution. The constitutional question for determination in this case cannot be stated as though it asks whether the distribution of electoral districts or of electoral regions is consistent with a general principle of representative democracy - especially if the content of "representative democracy" is derived from sources outside the Constitution. The constitutional question is whether there is inconsistency with the text and structure of the Constitution.
Unaffected by context, the phrase "chosen by the people" admits of different meanings. It might connote that candidates are chosen by popular direct election as distinct from election by an electoral college; or it might connote some requirement of equality or near equality of voting power among those who hold the franchise; or it might go further and import some requirement of a franchise that is held generally by all adults or all adult citizens unless there be substantial reasons for excluding them. Equally, these meanings might be attributed to the notion of "representative democracy". In this case we are not concerned with the mode of election: both the Council and the Assembly are elected by popular direct election. Nor are we concerned with the franchise. But the plaintiffs submit that an equality of voting power is implied in the Commonwealth Constitution.
Neither Nationwide News nor ACTV required an exhaustive determination of the content of the principle of representative democracy. Nor did the two defamation cases that followed, namely, Theophanous v Herald & Weekly Times Ltd [28] and Stephens v West Australian Newspapers Ltd [29] . All of these cases were concerned with the freedom of communication required to allow "the people" to perform their constitutional function of choosing their Parliamentary representatives. None of these cases was concerned with equality of voting power. The principle of representative democracy that was held to be implied in those cases can be no wider than - for it is synonymous with - what inheres in the text of the Constitution or in its structure. A submission that equality of voting power in State as well as in Federal elections is implied in the Constitution is not advanced by an appeal to the principle of representative democracy if reference to the text and structure of the Constitution fails to reveal an implication of that kind. However, if there be such an implication, an inconsistent State law would be invalid.
The invalidity would not flow from the operation of s 109: it would not depend on the enactment of an inconsistent law by the Parliament of the Commonwealth. The invalidity would arise because the power of the State Parliament to make (or to sustain) the law was withdrawn by the Commonwealth Constitution. The invalidity would arise from want of State legislative power. As the legislative power of a State is conferred by its Constitution and that Constitution is subject to the Commonwealth Constitution, the effect of any implication derived from the Commonwealth Constitution upon a purported law of a State must be ascertained by reference to the limitation placed by the Commonwealth Constitution on State legislative power.
The relationship between the Commonwealth Constitution and the Constitutions of the States
Section 106 of the Commonwealth Constitution reads:
The Constitution of each State of the Commonwealth shall, subject to this Constitution, continue as at the establishment of the Commonwealth, or as at the admission or establishment of the State, as the case may be, until altered in accordance with the Constitution of the State. This section has a dual operation. Its first operation is to prescribe what the new elements of the Federal polity - the States - shall be. When the people of the Australian Colonies were united in the Commonwealth of Australia by the proclamation made pursuant to Covering Cl 3 and those Colonies became Original States of the Commonwealth by operation of Covering Cl 6, the Colonies - the old constitutional entities - acquired a new constitutional status. They became States, as the text of ss 107 and 108 shows, deriving their existence as States from the Commonwealth Constitution [30] . Secondly, s 106 conferred on the respective States substantially the Constitutions of the antecedent Colonies [31] . The same Constitutions as had been conferred on the Colonies prior to 1 January 1901 were continued as the Constitutions of the respective States thereafter, subject to such modifications as were effected by the Commonwealth of Australia Constitution Act 1900 Imp and the Commonwealth Constitution. As Barwick CJ said in New South Wales v The Commonwealth [32] : On the passage of the Imperial Act, those colonies ceased to be such and became States forming part of the new Commonwealth. As States, they owe their existence to the Constitution which, by ss 106 and 107, provides their constitutions and powers referentially to the constitutions and powers which the former colonies enjoyed, including the power of alteration of those constitutions. Those constitutions and powers were to continue by virtue of the Commonwealth Constitution. 1. Victoria v The Commonwealth (1971) 122 CLR 353 at 371 2. China Ocean Shipping Co v South Australia (1979) 145 CLR 172 at 182 3. (1975) 135 CLR 337 at 372; see also Victoria v The Commonwealth (1971) 122 CLR 353 at 371-372 China Ocean Shipping Co v South Australia (1979) 145 CLR 172 at 182
Although the States owe their existence to the Commonwealth Constitution and although their respective Constitutions in 1901 were conferred by operation of s 106 of the Commonwealth Constitution, the powers vested in the respective State legislatures were subject not only to the other provisions of the Constitution but also to the limitations which had been imposed on the powers of the antecedent colonial legislatures (s 107). The Australian Colonies did not retain their colonial status for the purposes of the constitutional law of the Commonwealth, but their general legislative and constituent powers then remained amenable to affection by laws enacted by the United Kingdom Parliament [33] . That amenability to affection by Imperial legislation was removed only by the enactment of the Australia Act 1986.
In 1901, s 107 of the Constitution confirmed what s 106 might itself have provided, namely, that such power as a State Parliament possessed to alter its own Constitution [34] pursuant to the manner and form provisions of its Constitution and to alter those manner and form provisions themselves [35] remained [36] . Thus the "manner and form" provisions of the State Constitutions are retained as they were on 1 January 1901 or as they have been altered subsequently in accordance with any relevant manner and form provisions in force at the time of the alteration. It follows that the Constitution of a State at any time must be ascertained by reference to (i) its Constitution as at Federation; (ii) the overriding effect of the provisions of the Commonwealth of Australia Constitution Act and the Commonwealth Constitution; (iii) the modifications of the State Constitution that have been made either by Imperial legislation or State legislation provided, in the case of State legislation, it has been made in accordance with any relevant manner and form provisions of the particular State Constitution [37] ; and (iv) the Australia Act 1986. It is possible that a law of the Commonwealth could affect the Constitution of a State in some particular but not so as to curtail the continued existence of the State or the capacity of the Government of the State to exercise its functions [38] .
The Constitutions of the several States are, by force of s 106, subject to the Commonwealth Constitution, the provisions of which may be either expressed in its text or implied in its text and structure. There is no relevant reference in the Commonwealth Constitution to the distribution of the franchise in elections for State Parliaments. But the plaintiffs and the interveners supporting the plaintiffs submit that the Commonwealth Constitution implies a mandatory equality of voting power among the holders of the franchise for State as well as Commonwealth elections. Just as Mason CJ, Toohey and Gaudron JJ held in Stephens [39] that the concept of representative democracy is provided in the Constitution and is the basis of an implication of freedom of communication as to political matters relating to government at State level, so the concept of representative democracy is said to be the basis of an implication of equality of voting power in State elections. To consider this argument it is necessary, for reasons earlier stated, to take as the starting point not the concept of representative democracy but the text of the Commonwealth Constitution.
Equality of voting power: Commonwealth Constitution
The provisions of Pt III of Ch I of the Constitution and, in particular, s 24 were considered in McKinlay [40] . In that case, the chief question was whether equality of voting power was required in respect of elections for the House of Representatives. The Court answered this question in the negative. Barwick CJ found in the different franchises and electoral systems in force in the Australian colonies at the time of Federation an explanation of the terms of s 24. He said [41] :
[I]n my opinion, the expression "directly chosen by the people" is merely emphatic of two factors: first, that the election of members should be direct and not indirect as, for example, through an electoral college and, secondly, that it shall be a popular election. It is not an indirect reference to any particular theory of government. Members of the more numerous Houses of the Australian colonies, according to the franchises to which I have referred, were, in my opinion, elected directly by popular election. Within the meaning of the constitutional expression they could properly have been said to have been directly chosen by the people of the colony. Similarly, Gibbs J thought that the draftsmen of s 24 were "concerned with the manner of choice rather than with the people who were to choose" [42] . Mason J said [43] : Part III of Ch I of the Constitution contains several indications that equality or practically equality in the value of a vote, reflected in equality of numbers of electors or people represented by single member constituencies is not a constitutional requirement. Murphy J, in dissent, held that s 24 required an equal number of electors in each electoral division [44] . However, although McTiernan and Jacobs JJ rejected the contention that, in the light of history, s 24 should be understood to require that there be "absolute or as nearly as practicable absolute equality of numbers of the people" in an electorate, they acknowledged that the notion of equality is present, albeit remaining a matter of degree [45] . Stephen J adopted the same view [46] . 1. (1975) 135 CLR 1. 2. McKinlay (1975) 135 CLR 1 at 21 3. McKinlay (1975) 135 CLR 1 at 44 4. McKinlay (1975) 135 CLR 1 at 61 5. McKinlay (1975) 135 CLR 1 at 75 6. McKinlay (1975) 135 CLR 1 at 36-37 7. McKinlay (1975) 135 CLR 1 at 57cited above.
In Nationwide News [47] , there is a dictum by Deane and Toohey JJ [48] that tends in favour of the plaintiffs' argument:
While one can point to qualifications and exceptions, such as those concerned with the protection of the position of the less populous States [49] , the general effect of the Constitution is, at least since the adoption of full adult suffrage by all the States, that all citizens of the Commonwealth who are not under some special disability are entitled to share equally in the exercise of those ultimate powers of governmental control. However, the support for that proposition given by their Honours is a footnote referring generally to what McTiernan and Jacobs JJ had said in McKinlay [50] . 1. (1992) 177 CLR 1. 2. McKinlay (1992) 177 CLR 1 at 72 3. See, in particular, ss 7 (equal representation of the several Original States [in the Senate] shall be maintained) and 128 ("a majority of the States" as well as "a majority of all the electors"). 4. See, generally, McKinlay (1975) 135 CLR 1 at 36
Assuming, without deciding, that the provisions of the Commonwealth Constitution impliedly preclude electoral distributions that would produce disparities of voting power - of whatever magnitude - among those who hold the Commonwealth franchise in a State [51] , what do those provisions have to say with respect to the Constitutions and laws of the several States governing electoral distributions for State elections? In my opinion, the Commonwealth Constitution contains no implication affecting disparities of voting power among the holders of the franchise for the election of members of a State Parliament (hereafter "State disparities"). Far from containing an implication affecting State disparities, the text of Pts II and III of Ch I of the Commonwealth Constitution and the structure of the Constitution as a whole are inconsistent with such an implication. Sections 7, 8, 9 and 10 of the Constitution are expressed to relate only to elections for the Senate; ss 24, 25, 29 and 30 are expressed to relate only to elections for the House of Representatives; and s 41 is expressed to relate only to elections for either House of the Parliament of the Commonwealth. Not only are these sections expressly confined to elections for one or other House of the Commonwealth Parliament but they are all contained in Ch I of the Constitution which, being followed by Chs II and III, define the structure of the Commonwealth's three branches of government. Chapter V is the chapter relating to the States and their Constitutions. The structure of the Constitution is opposed to the notion that the provisions of Ch I might affect the Constitutions of the States to which Ch V is directed.
The steps in the plaintiffs' submission, so far as it is based on the Commonwealth Constitution, seem to be as follows: (i) the Commonwealth Constitution contains a principle of representative democracy; (ii) s 106 subjects State Constitutions to the Commonwealth Constitution; (iii) State Constitutions are therefore governed by the principle of representative democracy. But as the principle of representative democracy applies only to the process of electing members to either House of the Parliament of the Commonwealth and as the only provisions of the Commonwealth Constitution that are advanced as capable of affecting the State Constitutions relate to Federal elections, the submission fails at each step. The plaintiffs advance an alternative form of the submission, namely, that the principle of representative democracy implied in the text and structure of the Commonwealth Constitution informs the federal "organic unity of the Commonwealth and the States". I am not confident that I understand the submission. However it be understood, it must involve some notion of representative democracy controlling the distribution by State Parliaments of electoral power in relation to State elections. That proposition must be rejected. So long as the constitutionally implied principle is confined by the text and structure from which the implication is drawn, the only effect that it can have on the "organic unity" is in relation to the process of Federal elections. I would therefore reject the plaintiffs' submissions so far as they are based on the Commonwealth Constitution.
Equality of voting power: the Western Australian Constitution
In 1978, the Constitution Act 1889 WA (the 1889 Act) was amended [52] by inserting, inter alia, s 73(2). Sub-section (2) of s 73 prescribes that Bills of the several kinds therein mentioned should "not be presented for assent by or in the name of the Queen" unless passed by an absolute majority of the members of the Legislative Council and the Legislative Assembly and approved by electors at a referendum. Sub-section (2) thus "entrenches" those laws of Western Australia (including the Constitution Act) that would be affected by Bills of the kinds specified in that sub-section. The sub-section applies to a Bill of a kind specified in par (c), namely, a Bill that -
expressly or impliedly provides that the Legislative Council or the Legislative Assembly shall be composed of members other than members chosen directly by the people. In Stephens [53] , I said that s 73(2)(c) - entrenches in the Constitution Act the requirement that the Legislative Council and the Legislative Assembly be composed of members chosen directly by the people. This requirement is drawn in terms similar to those found in ss 7 and 24 of the Commonwealth Constitution from which the implication that effects a constitutional freedom to discuss government, governmental institutions and political matters is substantially derived. By parity of reasoning, a similar implication can be drawn from the Constitution Act with respect to the system of government of Western Australia therein prescribed. The "system of government" referred to is a Legislature "chosen directly by the people". It is one thing to hold that that prescription carries the implication in both the Commonwealth Constitution and the Western Australian Constitution that "the people" be free to discuss the matters that they need to discuss in order to choose their representatives; it is another thing to hold that the phrase implies that there be an equality of voting power. Whatever the phrase may prescribe in the Commonwealth Constitution in relation to the equality of voting power, its significance and effect in the Western Australian Constitution must be ascertained from its context and the circumstances in which s 73(2)(c) was enacted. 1. Acts Amendment (Constitution) Act 1978 WA (the 1978 Act), s 6. 2. (1994) 182 CLR 211 at 236.
Since the 1889 Act was assented to in 1890 pursuant to the Western Australian Constitution Act 1890 Imp, the electoral laws of Western Australia have been taken out of the 1889 Act where they were confirmed for the purposes of the then new Constitution of the State [54] . The laws relating to the distribution of electoral divisions were to be found in a number of different Acts in succession. In the 1899 Act, provision was made in ss 5 and 6 for Electoral Provinces for the election of the Council and in ss 18 and 19 for Electoral Districts for the election of the Assembly. Further distributions of Electoral Provinces and Electoral Districts were made by the Redistribution of Seats Acts of 1904, 1911 and 1929, the Electoral Districts Act 1922 and the Electoral Districts Act Amendment Act 1928. The 1947 Act introduced new provisions for the division of the State into Electoral Provinces for the election of the Council and Electoral Districts for the election of the Assembly. These provisions were themselves altered by the Electoral Districts Act Amendment Acts of 1955, 1963, 1965 and 1975. The last-mentioned Act (the 1975 Amendment) directed Electoral Commissioners to divide the Metropolitan Area into twenty-seven Electoral Districts and to divide the Agricultural, Mining and Pastoral Area into twenty-four Electoral Districts [55] , leaving the North-West-Murchison-Eyre Area with four Electoral Districts [56] . The "quota of electors in each area" [57] was to be taken as the basis for electoral districts, but a margin of allowance of 10 per cent in the Metropolitan Area and 15 per cent in the Agricultural, Mining and Pastoral Area was prescribed [58] . Sixteen Electoral Provinces for the election of the Council were to be distributed in stated numbers among the three Areas: the Metropolitan Area, the Agricultural, Mining and Pastoral Area and the North-West-Murchison-Eyre Area [59] . These were the provisions in force when s 73(2)(c) was inserted in the 1889 Act in 1978. In their operation these provisions authorised and perhaps required a disparity in voting power in respect of both elections for the Council and elections for the Assembly.
From before Federation, the State has been so divided into electoral districts for the election of the Assembly and electoral provinces or regions for the election of the Council that there has been inequality in the voting power of electors in different parts of the State. Throughout this time, the holders of the franchise as it stood from time to time have directly chosen the members of the Council and the Assembly for their respective electoral districts and provinces. When s 73(2)(c) was inserted in the 1889 Act in 1978, it is impossible to suppose that the Parliament of Western Australia intended thereby to override the regime of electoral districts and provinces which were then, and had historically been, the electoral framework of the State. The language of s 73(2)(c) must be construed in the light of the constitutional history of the State and the circumstances existing when that provision was introduced [60] . The purpose of that provision was not the creation of a new electoral regime; it was simply privative of the uncontrolled or partly-controlled power of constitutional amendment vested in the Parliament by s 73(1). The material purpose of the 1978 Act was expressed in its preamble as the making of further constitutional provision "to confirm the established constitutional provision aforesaid [relating to the Council, the Assembly and their powers] and to regulate the manner and form in which the powers of the Parliament of Western Australia may hereafter be exercised". Having regard to the history of the State's electoral laws and the context and operation of s 73(2)(c), it is impossible to find an implication other than the entrenchment of the system of electing members of the Council and the Assembly by direct popular vote. To find in s 73(2)(c) an implication that electoral power be equally distributed among the people of the State or among the people of the State possessing the franchise would be to find a legislative intention destructive of the means by which the enacting Parliament was elected.
I may add that, if the Western Australian Constitution had been affected by an implication drawn from the Commonwealth Constitution of the kind for which the plaintiffs contend, the provisions of the 1899 Act as in force on 1 January 1901 which then prescribed the Electoral Provinces for electing members of the Legislative Council (ss 5, 6) and the Electoral Districts for electing members of the Legislative Assembly (ss 18, 19) would have ceased to have effect. That would have been a startling and unintended consequence of Federation for Western Australia and probably for other States as well.
The legislative power of the Parliament of Western Australia to create or to alter or to authorise the creation or alteration of electoral divisions for the election of the Council and the election of the Assembly was unaffected by the insertion of s 73(2)(c) into the 1889 Act. There was no relevant amendment to the Western Australian Constitution after 1978 prior to the enactment of the 1987 Act. It follows that the Parliament had power to enact the impugned provisions of the 1987 Act. That Act does not touch the provisions, whether constitutional or statutory, that govern the election of members of the Senate or of the House of Representatives in the Commonwealth Parliament. The impugned provisions of the 1987 Act are therefore valid and the questions reserved for consideration by the case stated should be answered as follows:
(i) Is s 6 of the Constitution Acts Amendment Act 1899 WA invalid? Answer: No.
(ii) Are ss 2A(2), 6 and 9 of the Electoral Distribution Act 1947 WA, or any of them, invalid? Answer: No.
(iii) If any of ss 2A(2), 6 or 9 of the Electoral Distribution Act 1947 WA is or are invalid, is it or are they severable and, if so, to what extent? Answer: Unnecessary to answer.
In this case the plaintiffs contest the validity of Western Australian legislation which, in providing for the election of members of the Western Australian parliament, allows non-metropolitan electorates to comprise substantially fewer enrolled electors than metropolitan electorates. This, the plaintiffs say, offends against the principle which is succinctly, though inadequately, expressed in the political slogan "one vote, one value". Observance of that principle is, so the plaintiffs contend, required by implication by both the Commonwealth Constitution, which in that respect extends to Western Australia, and the Western Australian Constitution [61] .
For the purpose of the election of members of the Western Australian Legislative Assembly, the relevant legislation divides the State into two areas, one of which is the metropolitan area and the other of which is the rest of the State. Electoral distribution commissioners are required to divide the metropolitan area into thirty-four districts and the rest of the State into twenty-three districts [62] . The division into districts is required to be made in accordance with the principle that the number of enrolled electors comprised in any district must not be more than 15 per cent greater or less than a quotient obtained by dividing the total number of electors in the area by the number of districts into which the area is to be divided [63] . Since over 70 per cent of electors are enrolled in the metropolitan area, the effect of the legislation is that districts in the non-metropolitan area comprise substantially less electors than districts in the metropolitan area. Thus, as at 30 June 1995, the quotient for districts in the metropolitan area was 23,117 electors, whilst the quotient for districts outside the metropolitan area was 12,196.
Members of the Legislative Council are chosen from six regions, three of which are located in the metropolitan area. One of those three regions returns seven members and the other two return five members each. The remaining three regions are outside the metropolitan area. Two of those regions return five members each and the remaining one returns seven members [64] . The result is that less than 30 per cent of enrolled electors, being the electors outside the metropolitan area, elect 50 per cent of the members of the Legislative Council, leaving the other 50 per cent to be elected by over 70 per cent of the enrolled electors.
The argument that the Commonwealth Constitution requires electorates for the House of Representatives to contain, as far as practicable, equal numbers of persons or electors is not novel. Such an argument was put and rejected in Attorney-General (Cth); Ex rel McKinlay v The Commonwealth [65] . But the plaintiffs, encouraged by recent decisions of this Court [66] , put the argument once again, seeking leave, if necessary, to reopen McKinlay and to have the Court overrule that decision.
In the recent decisions upon which the plaintiffs rely, a majority of this Court found that ss 7 and 24 and related sections of the Constitution by implication require that there be freedom of communication in relation to political matters. Sections 7 and 24 provide that the members of the Commonwealth parliament be directly chosen by the people, in the case of the Senate by the people of each State and in the case of the House of Representatives by the people of the Commonwealth. For my own part, I was of the view in each of the decisions relied upon that ss 7 and 24 go no further than requiring that there be a choice of members of parliament by a direct, rather than an indirect, method. But the choice is to be made by the holding of elections (ss 10, 12, 29, 30) and is required to be a genuine choice so that, in the view which I expressed, any law which precluded the communication of information which was required for the exercise of a genuine choice would be invalid because it would be in conflict with ss 7 and 24 [67] .
The wider views expressed by other members of the Court were based upon the notion that the Constitution ordains representative government from which an implication of freedom of communication is to be drawn. Thus in Theophanous v Herald & Weekly Times Ltd [68] Mason CJ, Toohey and Gaudron JJ, having referred to Nationwide News Pty Ltd v Wills and Australian Capital Television Pty Ltd v The Commonwealth , said:
In those cases, a majority of the Court distilled from the provisions and structure of the Constitution, particularly from the concept of representative government which is enshrined in the Constitution, an implication of freedom of communication. That implication does not extend to freedom of expression generally [69] . The limited scope of the freedom was expressed in various ways by the members of the Court. It was described as "freedom of communication, at least in relation to public affairs and political discussion" [70] , "freedom to discuss governments and political matters" [71] , "freedom of communication about the government of the Commonwealth" which "extends to all political matters", including "matters relating to other levels of government" [72] , "freedom of political discourse" [73] and "freedom of participation, association and communication in relation to federal elections" [74] . Mason CJ, Toohey and Gaudron JJ went on in Theophanous [75] to say that it was necessary in considering the expression "political discussion" to bear in mind "that the underlying purpose of the freedom is to ensure the efficacious working of representative democracy". Deane and Toohey JJ had expressed a similar view in Australian Capital Television [76] where they said that "it is an implication of the doctrine of representative government embodied in the Commonwealth Constitution that there shall be freedom within the Commonwealth of communication about matters relating to the government of the Commonwealth". They continued: "the implication is drawn from an underlying doctrine of the Constitution rather than from any express term." 1. (1994) 182 CLR 104 at 120-121. 2. cf the First Amendment to the Constitution of the United States which reads: "Congress shall make no law abridging the freedom of speech, or of the press." 3. Australian Capital Television (1992) 177 CLR 106 at 138, 142, per Mason CJ. 4. Nationwide News (1992) 177 CLR 1 at 50, per Brennan J. 5. Australian Capital Television (1992) 177 CLR 106 at 169, per Deane and Toohey JJ. 6. Australian Capital Television (1992) 177 CLR 106 at 214, per Gaudron J. 7. Australian Capital Television (1992) 177 CLR 106 at 227, per McHugh J. 8. (1994) 182 CLR 104 at 123. 9. (1992) 177 CLR 106 at 168.
I have expressed in the previous cases the difficulty which I experience with this line of reasoning. It is a difficulty which, as I understand his judgment in Theophanous , McHugh J ultimately expressed when he said [77] :
I can find no support in the Constitution for an implication that the institution of representative government or representative democracy is part of the Constitution independently of the terms of ss 1, 7, 24, 30 and 41 of the Constitution. I think that all that can fairly be said is that those sections of the Constitution give effect to the political institution of representative government. But neither logic nor the efficacy of those sections or the federal system itself implies that independently of those sections the institution of representative government or representative democracy is itself part of the Constitution. The concept of representative democracy or representative government (and the latter is the more precise expression [78] ) does not have any necessary characteristics other than an irreducible minimum requirement that the people be "governed by representatives elected in free elections by those eligible to vote" [79] . Stephen J recognised as much in McKinlay when he observed that [80] : The principle of representative democracy does indeed predicate the enfranchisement of electors, the existence of an electoral system capable of giving effect to their selection of representatives and the bestowal of legislative functions upon the representatives thus selected. However the particular quality and character of the content of each one of these three ingredients of representative democracy, and there may well be others, is not fixed and precise. Sections 1, 7, 8, 16, 24 and 30 of the Constitution provide for the minimum requirements of representative government but do not purport to go significantly further. The Constitution also provides for the maintenance of equal representation of the Original States in the Senate and a minimum number of senators for each Original State (s 7), the rotation of senators (s 13), the filling of casual Senate vacancies (s 15), the disqualification of members (s 44), disputed elections (s 47) and certain other matters of machinery. It further provides in s 41 that no adult person who has or acquires a right to vote for the more numerous House of Parliament of a State shall, while the right continues, be prevented by any law of the Commonwealth from voting at elections for either House of the Parliament of the Commonwealth. Each elector has only one vote (ss 8, 30). Otherwise the form of representative government which we are to have is left to parliament, provision being made until parliament otherwise provides. (See ss 7, 8, 9, 16, 22, 24, 29, 30, 31, 34, 39, 46, 47, 48, 49.) In particular, it is left to parliament to make laws determining the electoral divisions for which members of parliament may be chosen. The only limitation is that a division shall not be formed out of parts of different States. (See ss 7, 9, 29.) In providing for those matters which are confided to it, parliament is required to determine questions of a political nature about which opinions may vary considerably. For example, the qualifications of electors are to be provided for by parliament under ss 8 and 30 and may amount to less than universal suffrage, however politically unacceptable that may be today. Thus, it may be seen that the form of representative government, including the type of electoral system, the adoption and size of electoral divisions, and the franchise are all left to parliament by the Constitution. 1. Theophanous (1994) 182 CLR 104 at 199 2. See Theophanous (1994) 182 CLR 104 at 189, per Dawson J; at 199-201, per McHugh J. 3. See Theophanous (1994) 182 CLR 104 at 201, per McHugh J; and Birch, Representative and Responsible Government: An Essay on the British Constitution (1964), p 15. 4. McKinlay (1975) 135 CLR 1 at 56
In those circumstances it was my view that no guarantee of freedom of communication, other than that required by the prescription of elections, could be read into the Constitution as an implication of representative government. The representative government prescribed by the Constitution is that for which it provides. The reasoning of the majority was to the contrary but that reasoning does not in my view extend so far as to support the plaintiffs' submission that, by implication, the Constitution requires electorates to be, as far as is practicable, of equal size.
There are hundreds of electoral systems in existence today by which a form of representative government might be achieved. Their merits must be judged by a number of different criteria which are likely to be incompatible with one another. As one commentator has put it [81] :
High priority amongst conflicting criteria would generally be given to such considerations as the extent to which a particular system promoted stable and effective government, fairness of representation, a wide choice of representatives, and contact between the electorate and its chosen representatives. But there will be disagreement on the relative priorities to be attached to each of these aims. Frequently a balance will have to be struck between them. Fair representation is a valuable aspiration, but not perhaps at the expense of encouraging the growth of too many splinter groups which could weaken the effectiveness of government. On the other hand, it would be foolish to pursue the aim of strong government so single-mindedly as to prevent the natural diversity of opinion amongst the electorate from being reflected in the composition of the legislature. There can be no implication that a particular electoral system, of the many available, is required by the Constitution. There is, of course, the express requirement that whatever system is employed it must result in a direct choice by the people. That must mean direct choice by the people through those eligible to vote at elections, but beyond that the matter of electoral systems, including the size of electoral divisions, and indeed whether to have divisional representation at all [82] , is left to the parliament. 1. Bogdanor, The People and the Party System (1981), p 209. 2. Constitution, s 29. See also Reeves and Ware, Electoral Systems: A Comparative and Theoretical Introduction (1992), pp 117-122.
Whatever those responsible for framing the Constitution may have regarded as the most appropriate electoral system for federal elections, their views are not contained in that instrument. They remain at best "unexpressed assumptions upon which the framers of the instrument supposedly proceeded" [83] and are not to be confused with those intentions which are expressed. Whilst implications can be and have been drawn from the Constitution, it is clear beyond question that implications may only properly be drawn where they are necessary or obvious. As Windeyer J observed in Victoria v The Commonwealth [84] "our avowed task is simply the revealing or uncovering of implications that are already there". That is to say, if implications are to be drawn, they must appear from the terms of the instrument itself and not from extrinsic circumstances [85] . The distinction has been drawn between textual and structural implications [86] , but I am not sure that the distinction is helpful. Whether or not an implication is categorised as structural or not, its existence must ultimately be drawn from the text. One is brought back to the text in the end and the danger in speaking of structural implications is, it seems to me, that there is a temptation to include by implication as part of the relevant structure those values which the structure is capable of accommodating, but does not necessarily accommodate [87] . That is, I think, what the plaintiffs' argument entails.
Once it is recognised, as in my view it must be, that electorates of equal numerical size are not a necessary characteristic of representative government, the plaintiffs are driven in their argument to find in the system of representative government laid down by the Constitution a requirement that there be, as nearly as practicable, electorates of equal size. But that requirement is nowhere to be found in any express provision of the Constitution and this Court has denied in McKinlay that there is any basis for its implication. It is not to be found in the expression "directly chosen by the people" contained in ss 7 and 24. Indeed, ss 7 and 24 contain requirements which are to the contrary [88] .
As Barwick CJ pointed out in McKinlay [89] no Australian colony at the time of federation insisted upon practical equality in the size of electoral divisions and the view was then plainly open that problems of communication and access in geographically large electorates outside a metropolitan area justify different numerical sizes in electoral divisions. That is a view which obviously still prevails in Western Australia under the current legislation.
Clearly there is force in the contrary view which holds that the effect of unequal electoral divisions - malapportionment - is to weight the value of votes in the numerically smaller divisions [90] . But the extra weight is only in the consequence that an elector in a smaller electorate is required to share his or her representative with a lesser number of electors than in the larger electorate. There are other ways, perhaps more significant, in which the value of a vote may be affected as, for example, where electoral divisions are defined in such a way as to allow one party in a two party system to return a majority of representatives with less than a majority of the total votes, which may occur whether or not malapportionment also exists. Disproportion of this kind may be intentionally caused by a gerrymander [91] . Of course, the problems arising from malapportionment and disproportion would largely disappear if there were no electoral divisions within a State and a system of proportional representation were adopted - something envisaged by s 29 of the Constitution [92] . But such a system may be to the detriment of a two party system by encouraging the growth of splinter groups. Barwick CJ adverted to the problems in McKinlay when, assuming electoral divisions, he said [93] :
Again, to ignore community of interest in the creation of electoral divisions and to insist on mere equality of numbers will be likely, in my opinion, to produce inequality rather than equality of voting value. It is probably impossible to devise a formula for electoral distribution which will necessarily produce equality in voting value, which will ensure that each vote is of equal weight in an election as a whole or even as between electoral divisions. 1. See Brugger and Jaensch, Australian Politics Theory and Practice (1985), pp 208-214 and Lijphart, Electoral Systems and Party Systems (1994), p 15. 2. See Bagger, The Supreme Court and Congressional Apportionment: Slippery Slope to Equal Representation Gerrymandering Rutgers Law Review, vol 38 (1985), pp 109, 109-110, 113, 116-118, 134-137; Davis v Bandemer (1986) 478 US 109 at 124, 127, 132; Shaw v Reno (1993) 125 LawEd 2d 511. 3. See Wesberry v Sanders (1964) 376 US 1 at 8 4. McKinlay (1975) 135 CLR 1 at 25
These considerations suggest that it would be unwise to freeze into a constitutional requirement a particular aspect of an electoral system the attraction of which might vary at different times, in different conditions and to different eyes. The wisdom of those who were responsible for framing our Constitution in recognising the political nature of such matters, and in leaving them to parliament, ought not to be overborne by drawing an implication which is neither apparent nor necessary.
In Reference re Electoral Boundaries Commission Act [94] the Canadian Supreme Court denied that the purpose of s 3 of the Canadian Charter of Rights and Freedoms was an absolute guarantee of equality of voting power. That section provides: "Every citizen of Canada has the right to vote in an election of members of the House of Commons or of a legislative assembly and to be qualified for membership therein." McLachlin J [95] held that effective representation, not equality of voting power per se, was the purpose of s 3 and that representation comprehends the idea of having a voice in the deliberations of government as well as the idea of the right to bring one's grievances and concerns to the attention of one's government representative. She continued [96] :
such relative [voter] parity as may be possible of achievement may prove undesirable because it has the effect of detracting from the primary goal of effective representation. Factors like geography, community history, community interests and minority representation may need to be taken into account to ensure that our legislative assemblies effectively represent the diversity of our social mosaic. These are but examples of considerations which may justify departure from absolute voter parity in the pursuit of more effective representation; the list is not closed. 1. (1991) 81 DLR (4th) 16. 2. Reference re Electoral Boundaries Commission Act [1991] 2 SCR 158; (1991) 81 DLR (4th) 16 La Forest, Gonthier, Stevenson and Iacobucci JJ concurring, Sopinka J agreeing with the result; Lamer CJ, L'Heureux-Dubé and Cory J dissenting. 3. Reference re Electoral Boundaries Commission Act [1991] 2 SCR 158 at 184; (1991) 81 DLR (4th) 16 at 36; cf MacKinnon v Prince Edward Island (1993) 101 DLR(4th) 362.
The Canadian Supreme Court rejected in Reference re Electoral Boundaries Commission Act the approach since 1964 in the United States which views the command in Art 1, s 2 of the United States Constitution that representatives be "chosen by the People of the several States" as a requirement that congressional electoral districts be as nearly as practicable equal in population [97] . That construction of Art 1, s 2 was adopted by the United States Supreme Court in Wesberry v Sanders [98] , having regard to the historical context of the provision, although in a forceful dissenting judgment Harlan J expressed the view that the "historical context" relied upon by the majority "bears little resemblance to the evidence found in the pages of history" [99] . However that may be, in Reference re Electoral Boundaries Commission Act McLachlin J distinguished the Canadian history of the right to vote in Canada from that of the United States and described the Canadian tradition as one of "evolutionary democracy moving in uneven steps toward the goal of universal suffrage and more effective representation, which even in its advanced stages tolerates deviation from voter parity in the interests of better representation" [100] .
The United States historical context is similarly inapt in any consideration of the form of representative government for which the Commonwealth Constitution provides. The democratic traditions of both Canada and Australia find their origins in the English model rather than in rebellion against it as is the case in the United States. Thus the reasoning in Wesberry v Sanders did not find favour with this Court in McKinlay [101] . As Mason J said [102] :
It is no answer to say that our Constitution is based on the American Constitution or that s 24 is based on Art 1, s 2. The interpretation adopted in Wesberry v Sanders is again a modern development that altered the inequality in electoral divisions which prevailed in the United States from early times. This inequality was understood to involve no contravention of Art 1, s 2 until Wesberry v Sanders was decided. When our Constitution was enacted it had not been decided that Art 1, s 2 insisted upon an equality in the value of a vote. The American provision was not thought to carry that message. And it is simply not correct to say that provisions in our Constitution should receive the same construction as that given to similarly worded provisions in the United States Constitution which have a different context and a different history, more particularly when the suggested construction is of recent origin, reversing an interpretation previously accepted. 1. (1975) 135 CLR 1 at 22-25, 45-47, 59, 62-63. 2. McKinlay (1975) 135 CLR 1 at 63
And as Gibbs J observed in McKinlay [103] the English model which was relevant in Australia at the time of federation did not display a history of electoral equality: after the Reform Acts and even after the Redistribution Act 1885 there remained substantial inequality.
Whilst the plaintiffs seek, if necessary, to have the Court overrule McKinlay , they attempt to sidestep that decision by placing reliance upon the recent decisions of this Court relating to freedom of communication and by finding in the concept of representative government not only an implied freedom of communication but also an implied requirement that electoral divisions be, as nearly as practicable, of the same numerical size. As I have indicated, the system of representative government which, without mentioning it by name, the Constitution prescribes is that for which it provides. And the provision which it makes is of a minimal kind. It is fallacious reasoning to posit a system of representative government for which the Constitution does not provide and to read the requirements of that system into the Constitution by implication. An implication of that kind is drawn from an extrinsic source and not from the text of the Constitution. It imports into the Constitution values which the Constitution does not adopt notwithstanding that it is capable of accommodating them. The Constitution does not, for these reasons, contain by implication the principle expressed in the words "one vote, one value", but the parliament may, should it consider it desirable to do so, adopt that principle in exercising its power to provide for electoral divisions. Indeed, it has done so in accordance with its view of the practicalities in the Commonwealth Electoral Act 1918 Cth [104] .
I would reject the plaintiffs' argument based upon the freedom of communication cases. That leaves the decision in McKinlay . In my view the reasoning of the majority in that case is compelling and I see no reason to depart from it.
It is true that in McKinlay McTiernan and Jacobs JJ [105] suggested that the notion of equality is to some extent present in the words "chosen by the people" so that at some point electoral inequality might be inconsistent with a choice by the people. They rejected, however, any requirement of absolute equality or nearly as practicable equality. Similarly, Mason J said [106] :
It is perhaps conceivable that variations in the numbers of electors or people in single member electorates could become so grossly disproportionate as to raise a question whether an election held on boundaries so drawn would produce a House of Representatives composed of members directly chosen by the people of the Commonwealth, but this is a matter quite removed from the proposition that s 24 insists upon a practical equality of people or electors in single member electorates. In my view, both McTiernan and Jacobs JJ and Mason J had in mind extreme situations markedly different from that which exists under the relevant Western Australian legislation. 1. McKinlay (1975) 135 CLR 1 at 36-37 2. McKinlay (1975) 135 CLR 1 at 61
Because of these conclusions it is strictly unnecessary for me to deal with the argument advanced by the plaintiffs that the requirement of electoral equality, said by them to be contained in the Commonwealth Constitution, extends to the States. But I am able to express my agreement with the conclusion reached by the Chief Justice, for the reasons given by him, that the system of representative government for which Ch I of the Commonwealth Constitution provides is confined to the Parliament of the Commonwealth and is not intended as a prescription for the States. The continuation by s 106 of the Commonwealth Constitution of the former colonial constitutions as State Constitutions is made "subject to this Constitution", but this does not serve to apply to the States provisions of the Commonwealth Constitution which otherwise have no application to them.
I am also in agreement with the Chief Justice, for the reasons which he gives, that, irrespective of the meaning to be given to the words "directly chosen by the people" in ss 7 and 24 of the Commonwealth Constitution, the words "chosen directly by the people" in s 73(2)(c) [107] of the Constitution Act 1889 WA do not contain any implication of electoral equality. In their own context they do not bear a meaning different from the similar words used in the Commonwealth Constitution.
For these reasons I agree with the answers proposed by the Chief Justice.
By this case stated the Court is asked to answer the following questions:
(i) Is s 6 of the Constitution Acts Amendment Act 1899 WA invalid?
(ii) Are ss 2A(2), 6 and 9 of the Electoral Distribution Act 1947 WA, or any of them, invalid?
(iii) If any of the ss 2A(2), 6 or 9 of the Electoral Distribution Act 1947 WA is or are invalid, is it or are they severable and, if so, to what extent?
Oversimplifying the plaintiffs' case, s 6 of the Constitution Acts Amendment Act 1899 WA (the 1899 Act) and ss 2A(2), 6 and 9 of the Electoral Distribution Act 1947 WA (the 1947 Act) are claimed to be invalid on the following grounds:
The basic contention of the plaintiffs is that by reason of one or more of the above considerations there is an implied requirement that, within practicable and rational limits, the number of persons eligible to vote for a member in one electorate of a House of the Parliament of the State of Western Australia should be approximately the same as the number of persons eligible to vote for a member in another electorate of that House. They say that, in respect of voting for members in both the Legislative Assembly and the Legislative Council, the legislation challenged results in an unwarranted malapportionment in favour of non-metropolitan voters.
The plaintiffs point to variations in the electoral districts of both the Legislative Assembly and the Legislative Council, which they say result in such a malapportionment between voters as to be at odds with the requirements of representative democracy. As a consequence of the 1947 Act, electoral districts for voting for the Legislative Assembly contain approximately 23,000 electors plus or minus 15 per cent for the metropolitan zone and approximately 12,000 electors plus or minus 15 per cent for the non-metropolitan zone. As to the Legislative Council, there are now six multi-member electoral regions, three in the metropolitan zone and three in the non-metropolitan zone. The distribution of seats is spread evenly between the two zones but the majority of electors is to be found in the metropolitan zone.
Thus the complaint of the plaintiffs is not that any one elector has a vote of greater value than the vote of any other elector in the constituency to which they belong. Rather, it is that there is an undue disparity in the number of electors who may vote in the various constituencies, the disparity being most marked as between metropolitan and non-metropolitan voters. It will be necessary to refer to this disparity in more detail later in this judgment.
The 1889 Act is expressed in its long title to be "An Act to confer a Constitution on Western Australia, and to grant a Civil list to Her Majesty". The 1899 Act is expressed in its long title to be "An Act to amend the Constitution Act 1889, and to amend and consolidate the Acts amending the same".
Section 6 of the 1899 Act [108] , which is under challenge by the plaintiffs, reads:
(1) The State shall be divided into 6 electoral regions under the Electoral Distribution Act 1947.
(2) The electoral regions known, respectively, as the North Metropolitan Region and the South West Region shall each return 7 members to serve in the Legislative Council.
(3) The electoral regions known, respectively, as the South Metropolitan Region, the East Metropolitan Region, the Agricultural Region and the Mining and Pastoral Region shall each return 5 members to serve in the Legislative Council.
Section 19 of the 1899 Act, which is not under challenge, provides that, for the purposes of the Legislative Assembly, the State shall be divided into fifty-seven electoral districts under the Electoral Act, "each returning one member to serve in the Legislative Assembly". 1. The section was substituted by the Acts Amendment (Electoral Reform) Act 1987 WA, No 40 of 1987, s 7.
The sections of the 1947 Act under challenge read [109] :
(1) The State shall be divided into districts and regions in accordance with this Act as soon as practicable after the day of the commencement of the Acts Amendment (Electoral Reform) Act 1987.
(2) If the same division under this Act has applied in respect of 2 successive general elections for the Legislative Assembly the State shall be divided into districts and regions in accordance with this Act as soon as practicable after the day that is one year after the polling day for the second of those general elections.
(3) The Governor may, by proclamation, direct that the State be divided into districts and regions in accordance with this Act as soon as practicable after the day of the issue of the proclamation.
(4) A proclamation shall be made under subsection (3) if both Houses of Parliament pass a resolution to that effect.
(5) The date used for determining the numbers of electors for the purpose of making a division required under subsection (1) or (2) or directed under subsection (3) shall be the day specified in subsection (1), (2) or (3) as the day as soon as practicable after which the division is to be carried out.
(a) divide the Metropolitan Area into 34 districts; and
(b) divide the area comprising the remainder of the State into 23 districts.
(2) The Commissioners shall make the division of an area mentioned in subsection (1)(a) or (b) into districts in accordance with the principle that the number of enrolled electors comprised in any district in the area must not be more than 15% greater, or more than 15% less, than the quotient obtained by dividing the total number of enrolled electors in the area by the number of districts into which the area is to be divided.
(a) 3 regions, to be known, respectively, as the North Metropolitan Region, the South Metropolitan Region and the East Metropolitan Region, each consist of complete and contiguous districts that together form the Metropolitan Area;
(b) one region, to be known as the Mining and Pastoral Region, consists of complete and contiguous districts that are remote from the capital and where the land use is primarily for mining and pastoral purposes;
(c) one region, known as the Agricultural Region, consists of complete and contiguous districts that together form an area that is generally south, or south and west, of and adjacent to the Mining and Pastoral Region; and
(d) the remaining region, to be known as the South West Region, consists of complete and contiguous districts.
Finally, it is necessary to set out a portion of s 73 of the 1889 Act, which is claimed by the plaintiffs to provide one basis upon which provisions of the 1947 Act should be declared invalid. So far as is material, s [110] reads:
(1) Subject to the succeeding provisions of this section, the Legislature of the Colony shall have full power and authority, from time to time, by any Act, to repeal or alter any of the provisions of this Act. Provided always, that it shall not be lawful to present to the Governor for Her Majesty's assent any Bill by which any change in the Constitution of the Legislative Council or of the Legislative Assembly shall be effected, unless the second and third readings of such Bill shall have been passed with the concurrence of an absolute majority of the whole number of the members for the time being of the Legislative Council and the Legislative Assembly respectively. Provided also, that every Bill which shall be so passed for the election of a Legislative Council at any date earlier than by Part III provided, and every Bill which shall interfere with the operation of sections 69, 70, 71, or 72, or of Schedules B, C, or D, or of this section, shall be reserved by the Governor for the signification of Her Majesty's pleasure thereon.
(a) expressly or impliedly provides for the abolition of or alteration in the office of Governor; or
(b) expressly or impliedly provides for the abolition of the Legislative Council or of the Legislative Assembly; or
(c) expressly or impliedly provides that the Legislative Council or the Legislative Assembly shall be composed of members other than members chosen directly by the people; or
(d) expressly or impliedly provides for a reduction in the numbers of the members of the Legislative Council or of the Legislative Assembly; or
(e) expressly or impliedly in any way affects any of the following sections of this Act, namely -
shall not be presented for assent by or in the name of the Queen unless -
(f) the second and third readings of the Bill shall have been passed with the concurrence of an absolute majority of the whole number of the members for the time being of the Legislative Council and the Legislative Assembly, respectively; and
(g) the Bill has also prior to such presentation been approved by the electors in accordance with this section,
and a Bill assented to consequent upon its presentation in contravention of this subsection shall be of no effect as an Act.
The preamble to the legislation adding sub-s (2) to s 73 reads in part:
And whereas by section 2 of the Constitution Act, 1889, it was provided that there shall be in place of the Legislative Council then subsisting a Legislative Council and a Legislative Assembly and that it shall be lawful for Her Majesty by and with the consent of the said Council and Assembly to make laws for the peace, order and good Government of Western Australia and its Dependencies and that such Council and Assembly shall, subject to the provisions of that Act have all the powers and functions of the then subsisting Legislative Council:
And whereas it is proper and expedient that further constitutional provision be made to further establish the offices aforesaid and to regulate the powers of the holders of those offices and to confirm the established constitutional provision aforesaid and to regulate the manner and form in which the powers of the Parliament of Western Australia may hereafter be exercised in relation to the offices and powers aforesaid.
Section 73(2)(c) requires that a Bill that "expressly or impliedly provides that the Legislative Council or the Legislative Assembly shall be composed of members other than members chosen directly by the people" shall not be presented for assent unless the second and third readings have been passed by an absolute majority of both Houses and the Bill has earlier been approved by the electors at a referendum.
In the course of the Second Reading Speech in the Legislative Assembly on the Bill that led to s 73 in its present form, the Premier said that it proposed that any future Bill "which would permit either House to be constituted by members not elected by the electors at large" would have to meet the double requirement of an absolute majority and approval at a referendum [111] . There is no reference in the Second Reading Speech in either House to the decision of this Court in Attorney-General (Cth); Ex rel McKinlay v The Commonwealth (McKinlay) [112] . However, given the relative shortness of time between judgment in McKinlay and the introduction of the amending legislation [113] , the legislature can hardly have been unmindful of the decision.
McKinlay concerned a provision of the Commonwealth Electoral Act 1918 Cth relating to any proposed distribution of a State into electoral divisions for elections to the House of Representatives. Section 24 of the Australian Constitution provides that the House of Representatives "shall be composed of members directly chosen by the people of the Commonwealth". By majority (Barwick CJ, McTiernan, Gibbs, Stephen, Mason and Jacobs JJ, Murphy J dissenting), the Court held that s 24 did not require the number of people or the number of electors in electoral divisions to be equal. The only difference in the language of s 24 of the Australian Constitution, "directly chosen by the people", and that of s 73(2)(c) of the 1889 Act, "chosen directly by the people", is one of juxtaposition. Therefore the reasoning in McKinlay has a direct bearing on the argument in the present case. Indeed the plaintiffs, while seeking to distinguish McKinlay , were driven as an alternative to asking the Court to overrule that decision. It will be necessary to look at McKinlay in greater detail; for the present it is enough to note that it stands in the way of the plaintiffs' argument that s 73(2)(c) of the 1889 Act itself demands equal value of voting power.
McKinlay may have a further significance. In Re Alcan Australia Ltd; Ex parte Federation of Industrial, Manufacturing and Engineering Employees [114] the Court said:
There is abundant authority for the proposition that where the Parliament repeats words which have been judicially construed, it is taken to have intended the words to bear the meaning already "judicially attributed to [them]" [115] although the validity of that proposition has been questioned [116] . It is true that McKinlay involved consideration of a provision of the Australian Constitution whereas s 73(2)(c) is an enactment of the legislature of Western Australia. Nevertheless the proposition mentioned in Alcan has some force in the present case. Certainly, in Burke v Western Australia [117] Burt CJ, with whom Smith J agreed, found the judgments of Barwick CJ, Gibbs J, Stephen J and Mason J supportive of the view of s 73(2)(c) that an electoral system whereby members are chosen by electors organised on a geographical basis satisfies the description of members "chosen directly by the people" and that this is so "irrespective of the number of people participating in that choice and irrespective of the ratio that that number may bear to the number participating in the choice of another member in and for another province or district as the case might be" [118] . 1. (1994) 181 CLR 96 at 106. 2. Barras v Aberdeen Steam Trawling & Fishing Co [1933] AC 402 at 446, per Lord Macmillan. See also D'Emden v Pedder (1904) 1 CLR 91 at 110 Pillar v Arthur (1912) 15 CLR 18 at 22, 25, 29-30; Platz v Osborne (1943) 68 CLR 133 at 141, 146, 146-147. 3. Salvation Army (Victoria) Property Trust v Shire of Fern Tree Gully (1952) 85 CLR 159 at 174, 182; R v Reynhoudt (1962) 107 CLR 381 at 388 Flaherty v Girgis (1987) 162 CLR 574 at 594 4. [1982] WAR 248. 5. Burke [1982] WAR 248 at 252
Legislative history and statutory construction tell against the plaintiffs' reliance on s 73(2)(c) in one way in which they seek to use it, that is, as directly requiring equality of voting power.
The Acts Amendment (Constitution) Bill (the First Bill) was introduced to the Parliament on 24 August 1977. The First Bill contained the equivalent to s 73(2)(c) in its present form. The First Bill was defeated in the Legislative Council on 19 October 1977 [119] . In the Second Reading Speech to the First Bill, the "right of electors at large to elect the members of either House" in the 1889 Act seems to be an underlying assumption [120] . However, the First Bill was not seen as making any substantive changes to the 1889 Act, rather as protecting and preserving the existing structures [121] . The purpose of the First Bill was to emphasise the role of the Queen in the Parliament, to preserve both Houses of Parliament, in particular to prevent the abolition of the Legislative Council, and to protect the office of Governor [122] . It was said not to affect the question of redistribution of electoral boundaries per se [123] and not to require a referendum to change them [124] .
At that time, there were considerable differences between the size of electorates [125] and the First Bill was not considered to be establishing the principle of one vote one value [126] . There was only a brief reference to the relevant aspect of cl 73 in the committee stage of the debate on the First Bill [127] and this reference suggests, if anything, that it was part of a scheme to prevent abolition of the Legislative Council. The second Acts Amendment (Constitution) Bill (the Second Bill) was in the same terms as the first [128] . It was also not seen as affecting the redistribution of electoral boundaries [129] .
The legislative history also suggests that s 73 is a manner and form provision. Certainly s 73(1) is a manner and form provision [130] and it has been read narrowly [131] . Sub-section (2) is also a manner and form provision, rather than any constitutional guarantee in itself. The preamble to the 1978 Act which added s 73(2) states that it is to "regulate the manner and form in which the power of the Parliament of Western Australia may hereafter be exercised".
A guide to the role of s 73(2)(c) in the present proceedings is to be found in a passage in the judgment of Mason CJ, Toohey and Gaudron JJ in Stephens v West Australian Newspapers Ltd (Stephens) [132] which reads:
We do not consider that s 73 provides a foundation for any suggestion that the Western Australian Constitution contemplates the possibility that it will be amended in such a way that representative democracy will be abolished. On the contrary, s 73(2) was plainly enacted with the object of reinforcing representative democracy and placing a further constitutional impediment in the way of any attempt to weaken representative democracy. And, so long, at least, as the Western Australian Constitution continues to provide for a representative democracy in which the members of the legislature are "directly chosen by the people", a freedom of communication must necessarily be implied in that Constitution. In other words, the whole of s 73(2) must be seen as a recognition that Western Australia is a representative democracy and as imposing a barrier against any attempt to detract from the elements of that polity as they are identified in the sub-section. 1. (1994) 182 CLR 211 at 233-234.
While the passage assumes importance in considering the inferences to be drawn from the 1899 Act itself, it does not carry the plaintiffs as far as they need to go in this aspect of their argument based on s 73(2)(c).
Whatever view be taken of the scope and operation of s 73(2)(c), there is a separate and wider question, namely, whether there is to be found in the 1889 Act a concept of representative democracy which requires an equality of voting power. Thus the statute has two aspects. One looks to s 73(2)(c) for a more specific dictate as to what is required in that respect. The other concerns the implications for representative democracy that may fairly be garnered from the Act as a whole.
To conclude that s 73(2)(c) does not of itself impose on the legislature of Western Australia an obligation to ensure that, within practicable and rational limits, the number of persons eligible to vote in one electorate should be the same as the number eligible to vote in another electorate does not dispose of the wider argument referred to in the preceding paragraph. In Burke v Western Australia [133] Burt CJ observed:
It may well be that the distribution of electors in the geographical area will be so unequal as to offend one's notion of fairness or to offend one's understanding of the idea conveyed by the words "representative democracy", but that is not to say that it offends against the idea of the sub-section. 1. [1982] WAR 248 at 252-253.
It is, I think, convenient to approach this part of the plaintiffs' argument by going first to the Australian Constitution to see whether the implication for which the plaintiffs contend is to be found there. If it is not, it is unlikely to be found in the Western Australian Constitution. If the implication is to be found in the Australian Constitution, a question then arises as to whether that implication thereby operates on the Western Australian Constitution and, if it does not, whether nevertheless such an implication is to be found in the State Constitution. If the implication is found in the Australian Constitution and thereby operates upon the Western Australian Constitution or if the implication is to be found in the latter, the plaintiffs have established the basic principle for which they contend.
The Australian Constitution: representative democracy
Recent decisions of the Court have held that the Australian Constitution prescribes a system of representative democracy or representative government. The terms have been used somewhat interchangeably but since it is representative democracy upon which the plaintiffs rely, I shall adhere to that term. The decisions which have given greatest emphasis to this prescription are Nationwide News Pty Ltd v Wills (Nationwide News) [134] , Australian Capital Television Pty Ltd v The Commonwealth (ACTV) [135] , Theophanous v Herald & Weekly Times Ltd (Theophanous) [136] and Stephens [137] . Those decisions necessarily left open the content of representative democracy and they did not touch the particular issues raised by the present application.
Different views have emerged as to the meaning and content of representative democracy. In ACTV [138] , Dawson J said that "much is left to the Parliament concerning the details of the electoral system to be employed in achieving representative democracy And, subject to the Constitution, the method of electing members of Parliament and the determination of electoral divisions also rest with the Parliament". However he continued that the Constitution provides for a true choice, in the sense of being an informed choice [139] . His Honour applied McKinlay as only requiring direct popular election [140] and saw the requirements of representative government in the Constitution as minimal [141] . McHugh J considered that the Constitution contains only a requirement of representative government, not the wider notion of representative democracy [142] . Representative government, he said, is "no more than a system under which the people were governed by representatives elected in free elections by those eligible to vote" [143] . McHugh J has taken the view that the conception of representative government at the time of federation is what is relevant [144] .
There are wider conceptions to be found in judgments of this Court. In ACTV Mason CJ said [145] : "The very concept of representative government and representative democracy signifies government by the people through their representatives." This holds the representatives accountable to the people for what they do [146] . Thus the people hold the ultimate sovereignty [147] . In Nationwide News, Deane J and I saw the election of members of Parliament as one way in which the ultimate power of governmental control reserved to the people under the Constitution was exercised [148] . We said [149] , adverting to what was said by McTiernan and Jacobs JJ in McKinlay [150] : "all citizens of the Commonwealth who are not under some special disability are entitled to share equally in the exercise of those ultimate powers of governmental control."
It is one thing to say that the Australian Constitution contains an implication of representative democracy. It is another to give content to that implication. The Court has said that freedom of communication on political matters is an essential requirement of representative democracy. Are there other essential requirements and, in particular, is there one that is relevant to the present case? Clearly, no implication can be drawn in absolute terms. Nor do the plaintiffs seek to do so. Their claim is for a requirement of electoral equality "within practicable and rational limits".
The Australian Constitution does not use the expression representative democracy or, indeed, representative government. Because democracy is a "dynamic phenomenon" [151] , its significance within the Constitution cannot be frozen by reference to the year 1900 or thereabouts. The Constitution must be construed as a living force [152] and the Court must take account of political, social and economic developments since that time. Even before 1900, Garran said that responsible government "as we know it, is a new thing and a changing thing" [153] . And in 1908, though in a different context, O'Connor J said that "it must always be remembered that we are interpreting a Constitution broad and general in its terms, intended to apply to the varying conditions which the development of our community must involve" [154] . Similarly, Dixon J said [155] : "it is a Constitution we are interpreting, an instrument of government meant to endure and conferring powers expressed in general propositions wide enough to be capable of flexible application to changing circumstances."
Another consideration is that whilst the connotation of words in the Australian Constitution remains fixed, their denotation may vary over time [156] . Equally, while the "essential feature" [157] of a concept in the Constitution remains unchanged, the concept is necessarily applied to circumstances different to those which existed at the beginning of this century. The connotation/denotation distinction has traditionally been applied where there have been technological advances [158] . However in Cheatle v The Queen , a unanimous Court took a somewhat different approach in interpreting the phrase "trial by jury" in s 80 of the Constitution. The Court distilled the representative character of a jury as one of the "essential features" of trial by jury [159] . But it noted that the qualifications of jurors had changed since 1900 when only men who satisfied a property qualification were eligible. The Court said [160] : "The exclusion of women and unpropertied persons was, presumably, seen as justified in earlier days by a then current perception that the only true representatives of the wider community were men of property." It went on to hold that, in 1993, representative juries under s 80 required that women and unpropertied persons be also eligible as jurors [161] .
By parity of reasoning, the essential feature of representative democracy is government by the people through their representatives. "[T]he powers of government belong to, and are derived from, the governed, that is to say, the people" [162] . In 1900, the popular perception of what this entailed was certainly different to current perceptions. For instance, the franchise did not include all, or even a majority, of the population [163] . But according to today's standards, a system which denied universal adult franchise would fall short of a basic requirement of representative democracy [164] . The point is that, while the essence of representative democracy remains unchanged, the method of giving expression to the concept varies over time and according to changes in society. It is the current perception which is embodied in the Australian Constitution. The point was made by McTiernan and Jacobs JJ in McKinlay when they said of s 24 of the Constitution [165] :
The words "chosen by the people of the Commonwealth" fall to be applied to different circumstances at different times and at any particular time the facts and circumstances may show that some or all members are not, or would not in the event of an election, be chosen by the people within the meaning of these words in s 24. At some point choice by electors could cease to be able to be described as a choice by the people of the Commonwealth. It is a question of degree. 1. Nationwide News (1992) 177 CLR 1 at 72, per Deane and Toohey JJ. 2. The franchise varied from State to State, but women, Aboriginals and persons of particular racial groups, non-propertied persons, non-professional persons or those without university degrees were amongst those excluded from the franchise. For a discussion of State franchises, see McKinlay (1975) 135 CLR 1 at 19, per Barwick CJ. 3. McKinlay (1975) 135 CLR 1 at 36, per McTiernan and Jacobs JJ; at 69, per Murphy J. 4. McKinlay (1975) 135 CLR 1 at 36
Likewise, around 1900 the method of choosing representatives involved significant inequality in voting power. At that time, the ratio between the number of electors in the largest and smallest electorates varied from 2:1 for the Legislative Council of South Australia to 38:1 for the Legislative Assembly of Western Australia. Clearly the expression of representative democracy then did not encompass equality of voting power. But just as clearly, the expression of the concept is now thought to do so, at least in this country. Equality of electorate size, with an allowable percentage variation, is now required for parliamentary elections in the Commonwealth [166] , New South Wales [167] , Victoria [168] , Queensland [169] , South Australia [170] and Tasmania [171] . This move towards equality of electorate size reflects a change in society's perception of the appropriate expression of the concept of representative democracy. As Mason J recognised in McKinlay [172] : "the conception of equality in the value of a vote or equality as between electoral divisions is a comparatively modern development." In its final report the Constitutional Commission concluded [173] : "We believe one vote one value is an essential principle of democracy. It is fundamental to a sense of meaningful participation in Australia's democratic polity."
Once it appears, as recent decisions of the Court make clear, that the Australian Constitution established a system of representative democracy, it is apparent that more is involved than freedom of political communication. In Australian National Airways Pty Ltd v The Commonwealth [174] Dixon J advised against "pedantic and narrow constructions in dealing with an instrument of government", saying: "I do not see why we should be fearful about making implications." To this Windeyer J added the rider [175] : "I would prefer not to say "making implications", because our avowed task is simply the revealing or uncovering of implications that are already there."
The plaintiffs placed reliance on decisions of the United States Supreme Court which have held that, as nearly as practicable, one person's vote should be worth as much as another person's [176] . However these decisions are based on the intentions of the framers of the American Constitution and have been influenced by the history of that country. For these reasons, several of the Justices in McKinlay [177] held that decisions of that Court provided little guidance in this area of the Australian Constitution. I respectfully agree [178] .
On the other hand, decisions of Canadian courts are more relevant as, like Australia, Canada "adopted and built on the English tradition" [179] . Canadian courts have recognised that the notion of equality is fundamental to the right to vote guaranteed by s 3 of the Canadian Charter of Rights and Freedoms [180] . As Cory J (dissenting), with whom Lamer CJ and L'Heureux-Dubé J agreed, said in Reference re Electoral Boundaries Commission Act [181] :
[T]he right to vote is fundamental to a democracy. If the right to vote is to be of true significance to the individual voter, each person's vote should, subject only to reasonable variations for geographic and community interests, be as nearly as possible equal to the vote of any other voter residing in any other constituency. Similarly, McLachlin J, with whom La Forest, Gonthier, Stevenson and Iacobucci JJ agreed, said [182] : What are the conditions of effective representation? The first is relative parity of voting power. A system which dilutes one citizen's vote unduly as compared with another citizen's vote runs the risk of providing inadequate representation to the citizen whose vote is diluted. 1. Dixon v Attorney-General (British Columbia) (1989) 59 DLR(4th) 247 at 260, per McLachlin CJ. 2. Dixon v Attorney-General (British Columbia) (1989) 59 DLR(4th) 247; Reference re Electoral Boundaries Commission Act [1991] 2 SCR 158; (1991) 81 DLR (4th) 16; Reference re Electoral Boundaries Commission Act (Alberta) (1991) 86 DLR (4th) 447; MacKinnon v Prince Edward Island (1993) 101 DLR(4th) 362; Reference re Electoral Divisions Statutes Amendment Act, 1993 (Alberta) (1994) 119 DLR (4th) 1. 3. (1991) 81 DLR (4th) 16 at 26. 4. Reference re Electoral Boundaries Commission Act (1991) 81 DLR (4th) 16 at 35.
The principle thus enunciated is not in absolute terms but it follows that a general principle of equal electorates is "a minimal requirement for a representative democracy" [183] . In that event the principle is part of the Australian Constitution, even if not expressed in any provision. It derives from s 24 (chosen by the people) but more fundamentally from the very structure of the system of government enshrined in the Constitution. In the end it must be accepted that equality of political rights is not achieved "if the vote of a person in one part of the country has a greater effect in securing parliamentary representation than the vote of a person in another part of the country" [184] . The principle cannot be stated in absolute terms because we are speaking of a representative democracy in which each citizen is entitled to be represented in government. The goal is effective representation of all citizens. That precludes absolute parity not only as a matter of practicalities but also for the reason expressed by McLachlin J in the following terms [185] :
Factors like geography, community history, community interests and minority representation may need to be taken into account to ensure that our legislative assemblies effectively represent the diversity of our social mosaic. Nevertheless, recognition of these considerations does not detract from the fundamental importance to democracy of the principle itself. 1. Creighton, Apportioning Electoral Districts in a Representative Democracy University of Western Australia Law Review, vol 24 (1994) 78, at p 83. 2. O'Donovan v Attorney-General [1961] IR 114 at 137 3. Reference re Electoral Boundaries Commission Act [1991] 2 SCR 158 at 184; (1991) 81 DLR (4th) 16 at 36.
Equality of voting power is an underlying general requirement in the Constitution. To a limited extent it has given way to the requirements for a minimum of five members of the House of Representatives for each State (s 24) and for equal numbers of senators for each State (s 7). The defendant argues that these provisions tell against the principle claimed by the plaintiffs.
This argument can be rejected in relation to the Senate. The Senate is "the chamber in which the States, considered as separate entities, and corporate parts of the Commonwealth, are represented" and "[t]hat the States, and not the people, are actually represented in the Senate is shown by the requirement that the "equal representation of the several Original States shall be maintained" " [186] .
The argument can also be rejected in relation to the House of Representatives. The provision for a minimum of five members was seen by Barwick CJ and Gibbs J in McKinlay [187] to preclude voting equality. But it is an historical concession to Tasmania and Western Australia which would otherwise have been entitled to only two or three members. This was considered to be an insignificant representation [188] . Other aspects of s 24 support an implication of equality of voting power [189] . In particular, s 24 embraces proportional representation as a general principle; the minimum of five members is merely an exception to this and today it only applies to Tasmania.
As noted earlier, the decision in McKinlay is relied upon by the defendant as an answer to the plaintiffs' case. It is true that the Court (Barwick CJ, McTiernan, Gibbs, Stephen, Mason and Jacobs JJ, Murphy J dissenting) held that s 24 did not require the number of people or the number of electors in electoral divisions under the Commonwealth Electoral Act 1918 Cth to be equal. But in my view it is not necessary for the Court to overrule McKinlay in order to accede to the plaintiffs' general argument. McKinlay can be distinguished from the present case.
One suggested basis for doing so is that in McKinlay the plaintiff's argument was for a requirement of absolute equality tempered only by the demands of practicability and that the decision would not be inconsistent with a requirement for a lesser measure of equality, qualified by other considerations [190] . But in McKinlay the plaintiffs' argument was that s 24 demanded, as far as practicable, equality of electorate size [191] ; they were not contending for absolute equality. Even if this distinction was valid, in the present case the plaintiffs couched their claim in terms of a requirement that, within practicable and rational limits, the number of persons eligible to vote for a member in one electorate should be approximately the same as in another electorate. There is no meaningful distinction between the principle claimed in either case. Although the current plaintiffs accepted that other considerations might intrude, these would constitute a qualification on the principle of practical equality.
An important point of distinction is that in McKinlay the Court was examining whether s 24, in particular the phrase "directly chosen by the people", required equality of voting power. Most members of the Court did not examine the requirements of representative democracy. Only Gibbs J and Stephen J looked at the wider question of whether such equality was a necessary aspect of a representative democracy [192] . So while the majority of the Court held that s 24 did not contain a requirement of voting equality, there is no majority decision on whether such a requirement is to be found in the notion of representative democracy which is the underlying basis of the Constitution. In that respect it is important that the decision preceded the recent decisions of this Court concerning representative democracy. It is unnecessary to overrule McKinlay in order to accede to the proposition that a general principle of equality of voting power is an aspect of the Australian Constitution.
The plaintiffs argue that the organic unity of the Commonwealth and States means that they are inexorably entwined so that an implication of electoral equality in the Australian Constitution will necessarily apply to Western Australia. In Nationwide News Deane J and I spoke of the Constitution's doctrine of representative government as "structured upon an assumption of representative government within the States" [193] . It was unnecessary to explore the implications of this assumption other than in relation to the implied freedom of political discussion.
Some judgments of the Court have treated the implied freedom of political discussion in the Australian Constitution as necessitating a similar freedom at State level. The reason is to be found in the indivisible nature of public discussion and the inter-relationship between the various tiers of government [194] . But it does not follow that, in regard to the electoral process, an implication of equality in the Commonwealth necessitates such an implication in the States.
There are sections of the Australian Constitution that make State electoral laws applicable to Commonwealth elections (ss 7, 9, 10, 15, 25, 29, 30, 31, 41). Most of these provisions are transitional and permit State electoral laws to be overridden by Commonwealth laws (ss 7, 9, 10, 29-30). Section 15, which deals with casual vacancies, comes closest to making a direct connection between the parliaments of the Commonwealth and the States [195] .
Various submissions were presented to the Court by reference to s 15. The plaintiffs argued that the section presupposes that the States will have a system which in a general way corresponds with the Commonwealth system. The Commonwealth argued that the section reflected the assumption that State parliaments are predicated on the same basis as the Commonwealth Parliament. South Australia argued that it is not logically or practically necessary for the preservation of representative democracy at the Commonwealth level that there be representative democracy in the States. Tasmania pointed out that even if State electoral laws did not ensure representative democracy, this would not necessarily affect the working of the Commonwealth Parliament and in any case, the Commonwealth could insulate itself from such laws [196] .
Section 15 has not been interpreted as imposing restrictions upon State parliaments. In particular, it has been held that it does not preclude the abolition of one of the Houses of Parliament of a State [197] . In Clayton v Heffron , Dixon CJ, McTiernan, Taylor and Windeyer JJ said [198] :
The supposition that there will be two Houses implies no intention legislatively to provide that the constitutional power of the State to change to a unicameral system, if the power existed, should cease. One can understand the section being relied upon as evidence that it was not supposed that the power to make the change existed. But that is all. Even that is not a very cogent argument. This passage was relied on by McHugh J in Theophanous to deny any connection between a requirement of representative democracy in the Commonwealth Parliament and the State parliaments [199] . But, in my view, that is to draw too much from the observations of their Honours which were directed to a different issue. 1. Clayton v Attorney-General (NSW) [1960] NSWR 592, at 618-619; Clayton v Heffron [1961] SR(NSW) 768 at 821, 833, 855; Clayton v Heffron (1960) 105 CLR 214 at 248-249 2. Clayton v Heffron (1960) 105 CLR 214 at 248-249 3. Clayton v Heffron (1994) 182 CLR 104 at 202
It is difficult to justify giving s 15 the role contended for by the plaintiffs, given its remedial nature. Even if the replacement senator was chosen by a State parliament which was not a representative government, the effect on representative government at the Commonwealth level would be minimal. As Isaacs J said in Vardon v O'Loghlin [200] :
[Section] 15 of the Constitution was not framed with the object of meeting numerous instances of irregular Senate elections, but of providing for possible but rare contingencies of the abnormal termination of the service of senators; so rare that departures from the fundamental principles of representation through popular election would be really inappreciable because infrequent and possibly of short duration. 1. (1907) 5 CLR 201 at 216.
The plaintiffs argue that the system of representative democracy at the Commonwealth level imposes restrictions on State legislative powers through the operation of s 106 of the Australian Constitution.
Section 106 is in Ch V of the Constitution - "The States". It reads:
The Constitution of each State of the Commonwealth shall, subject to this Constitution, continue as at the establishment of the Commonwealth, or as at the admission or establishment of the State, as the case may be, until altered in accordance with the Constitution of the State. The "States" are identified in covering cl 6 of the Commonwealth of Australia Constitution Act 1900 Imp as such of the colonies there referred to "as may be admitted into or established by the Commonwealth as States".
It has been said that the provision "has always raised a number of difficult questions concerning the relationship between it and the State Constitutions which it sought to protect, and between those Constitutions and the Commonwealth Constitution as a whole" [201] . In the course of discussing s 106, Quick and Garran comment [202] :
The States retain their executive, legislative, and judicial departments as before, but shorn of some of their powers and functions. Certainly the States derive their existence as States from the Australian Constitution [203] . The original source of authority for the legislative power of the States is Imperial legislation but the present source of that power is s 106 [204] . Even so, it does not follow that the plaintiffs' submission is correct. 1. Craven, A Few Fragments of State Constitutional Law, University of Western Australia Law Review, vol 20 (1990) 353, at p 367. 2. Quick and Garran, The Annotated Constitution of the Australian Commonwealth (1901), p 931. 3. Victoria v The Commonwealth (1971) 122 CLR 353 at 371 New South Wales v The Commonwealth (1975) 135 CLR 337 at 372 4. Victoria v The Commonwealth (1971) 122 CLR 353 at 371-372 New South Wales v The Commonwealth (1975) 135 CLR 337 at 372 China Ocean Shipping Co v South Australia (1979) 145 CLR 172 at 182 The Commonwealth v Queensland (1975) 134 CLR 298 at 336-337 Bistricic v Rokov (1976) 135 CLR 552 at 566 Western Australia v Wilsmore (1982) 149 CLR 79 at 86cf Western Australia v Wilsmore [1981] WAR 179 at 182-183
By reason of s 106, the "Constitution of each State" continues "as at the establishment of the Commonwealth", until altered in accordance with the Constitution of the State, subject, however, to "this Constitution" [205] . The identification of this Constitution in s 106 is clear enough. But the scope and operation of s 106 are by no means settled. Historically it has been seen, for the most part, as offering protection to the States against the exercise of Commonwealth power. Thus in Australian Railways Union v Victorian Railways Commissioners Dixon J expressed the matter more cautiously than other members of the Court when he said [206] :
It may be that s 106 provides the restraint upon the legislative power over States which differentiates it from the power over the subject and that no law of the Commonwealth can impair or affect the Constitution of a State. 1. Whatever effect the laws of the United Kingdom Parliament might have continued to have on the former colonies was removed by the Australia Act 1986. 2. Australian Railways Union (1930) 44 CLR 319 at 391-392.
Section 106 played a significant part in the decision of the Court in Re Tracey; Ex parte Ryan [207] that certain provisions of the Defence Force Discipline Act 1982 Cth were invalid. Other decisions treating s 106 as protective of State Constitutions are Attorney-General (NSW) v Ray [208] and S (a Child) v The Queen [209] .
The plaintiffs argue that s 106 makes the constitution of a State subject to the Australian Constitution and that therefore, to the extent that an implication of representative democracy is to be found in the latter, it will operate in respect of the former. Prior to Theophanous , the only support for giving s 106 such an operation was from Murphy J in Western Australia v Wilsmore [210] in holding that a contravention of a State Constitution was a contravention of the Australian Constitution. However this does not really support the plaintiffs' argument as Murphy J was dealing with a manner and form requirement in a State Constitution; in that respect s 106 asserts the need to comply with the Constitution of the State [211] .
In Theophanous there are suggestions that the implied freedom of political communication limits the legislative powers of the States on this aspect through the operation of s 106 of the Australian Constitution [212] . The strongest support comes from Deane J who treated s 107 of the Constitution, which continues the legislative powers of the State parliaments, as a provision to be read with s 106 so that those powers are subject to the Constitution. He placed emphasis on the provision in s 106 that State Constitutions continued "subject to this Constitution" and held that State legislative powers were restricted by the freedom of political communication guaranteed by the Australian Constitution.
But even if emphasis is placed on the words "subject to this Constitution", what consequence does this have for State Constitutions in relation to equality of voting power as an aspect of representative democracy contained in the Australian Constitution? The words "subject to this Constitution" override a particular aspect of a State Constitution in so far as it is contrary to the Australian Constitution [213] . The usual application of this is in the context of State legislative powers, so that State powers are to give way to the requirements of the Australian Constitution [214] .
But it does not follow that an implication of equality of voting power at the Commonwealth level effects, through s 106, an implication of equality of voting power at the State level. The words "subject to" mean "if not inconsistent with or repugnant to" [215] . Any guarantee of voting equality in Commonwealth elections will not be affected by State electoral laws permitting inequality in State elections. In this respect there is no necessary inconsistency between voting inequality at the State level and voting equality at the Commonwealth level. The conduct of State elections will not undermine Commonwealth elections. The implication of freedom of political communication can be distinguished, for it is the nature of such communication that State restrictions may undermine the Commonwealth guarantee. This was a significant factor in Deane J's reasoning in Theophanous [216] . Thus while Deane J relied on s 106 to find that State legislative powers were subject to the implied freedom of political communication, his reasoning does not mean that any Commonwealth guarantee of voting equality translates into a State guarantee of voting equality. Of course, if a provision of a State Constitution purported to affect the Commonwealth guarantee of voting equality, it would be overridden through the operation of s 106.
Section 106 does not effect a blanket importation of the Australian Constitution into State Constitutions. To interpret s 106 in this way unduly subjects State Constitutions to the Australian Constitution at the price of the other stated aims of the section. Its primary aim is to guarantee the continuation of State Constitutions after federation [217] , though subject to the Constitution.
Implication of representative democracy in the 1889 Act
However, the concept of representative democracy is to be found in the Constitution of Western Australian. It exists by reason of its similarity in relevant respects to the Australian Constitution and by reason of the elements of representative democracy recognised in s 73(2) of the 1889 Act. The elements in the Australian Constitution which lead to the conclusion that it is predicated on a system of representative democracy are present in the 1889 Act. One element is the vesting of legislative power in the Parliament, consisting of the Queen, the Senate and the House of Representatives [218] . This element is also present in the 1889 Act [219] . Another relevant factor is the requirement of approval by the people for the amendment of the Australian Constitution [220] . In the 1889 Act, there is a requirement of electoral approval, though it does not cover all amendments [221] . In the recent judgments of the Court to which reference has been made reliance was placed upon the control given to the people over the composition of Parliament by ss 7 and 24 [222] .
In Stephens [223] a majority of the Court found that representative democracy was contained in the Western Australian Constitution for the same reasons that it was contained in the Australian Constitution. Emphasis was placed on the phrase "chosen directly by the people" in s 73(2)(c) of the 1889 Act in a provision that "was plainly enacted with the object of reinforcing representative democracy and placing a further constitutional impediment in the way of any attempt to weaken representative democracy" [224] .
One argument to the contrary is that no implications can be drawn from State Constitutions because they are subject to amendment. Taken to its extreme, this argument denies the drawing of implications even from entrenched provisions because they can still be changed, even though change is more difficult than for non-entrenched provisions. This argument must be rejected in the light of the decisions of this Court which have found implications in the Australian Constitution, an instrument which can be changed, albeit only by the method laid down in s 128. More generally, implications are drawn from constitutions as they presently stand. It is not to the point that these constitutions may be changed in the future so that the implication ceases to have a valid basis.
The more limited version of this argument is that implications from State Constitutions can only be drawn from entrenched provisions. This is subject to the same response just given. But it may have more merit in so far as any implication drawn from non-entrenched provisions may be overridden by legislative amendment. On the other hand, even entrenched provisions may be amended. The validity of this argument does not need to be decided for the purposes of the present case as the relevant provisions of the 1889 Act in which representative democracy can be found are entrenched. Section 73(2) entrenches provisions providing for the Governor, Legislative Council and Legislative Assembly and manner and form requirements.
A further argument is that even if an implication is drawn from a State Constitution, it would not have the status to impose a restriction on the plenary legislative powers of the State. This misses the point made in the cases in which an implied freedom of political communication was found in the Australian Constitution. These cases show that the implication operates as an implied restriction and control on legislative powers, not just as an abstract principle.
It is one thing to enunciate equality of voting power as an aspect of representative democracy. It is another to determine whether a particular electoral system is at odds with that principle. It is yet another thing to identify the limits of justiciability in such a situation. What is under challenge here is the validity, indeed the constitutionality, of certain legislative provisions. There can be no question of non-justiciability in that event though the Court must give due recognition that the principle said to have been offended by the provisions in question cannot be asserted in absolute terms and therefore much must be left to the political judgment of the legislature.
Mention has been made of the observation of Burt CJ in Burke v Western Australia [225] that the distribution of electors may be so unequal as to offend one's understanding of representative democracy. And in McKinlay [226] Mason J said:
It is perhaps conceivable that variations in the number of electors or people in single member electorates could become so grossly disproportionate as to raise a question whether an election held on boundaries so drawn would produce a House of Representatives composed of members directly chosen by the people of the Commonwealth. 1. [1982] WAR 248 at 252-253. 2. (1975) 135 CLR 1 at 61. See also at 57, per Stephen J.
Unlike the United States, the Canadian courts appear to have baulked at identifying a ratio beyond which there is such an imbalance as to invalidate the electoral system. It is true that in Dixon v Attorney-General (British Columbia) [227] McLachlin CJSC said:
[I]t is appropriate to set limits beyond which [equality of voting power] cannot be eroded by giving preference to other factors and considerations, such as the 25 per cent limit applied in Canada to federal electoral districts or the 10 per cent limit established recently in Australia. Nevertheless, she did not identify the limits which might be set. 1. (1989) 59 DLR (4th) 247 at 266-267.
The courts must exercise restraint in this area, in particular not seeking in effect to say how boundaries should be drawn or prescribing specific ratios that are acceptable. Clearly there may be room for disagreement as to whether a particular system offends the principle of electoral equality demanded by representative democracy. Indeed the plaintiffs speak of an implied requirement "within practicable and rational limits". It may be said that if the courts do not prescribe limits, the task of legislatures is made more difficult and the scope for challenging an electoral system is inevitably widened. The answer is that an approach in broad terms gives proper recognition to the respective roles of the legislature and the court. The task of the court is to identify and give effect to the constitutional principle at issue.
As a result of the legislation now in force, the Legislative Assembly comprises fifty-seven members (s 18(b) of the 1899 Act). Each member is elected for a single district (s 19(b) of the 1899 Act). Of the fifty-seven members, thirty-four are returned for districts in the Metropolitan Area [228] and twenty-three for the remainder of the State (s 6(1) of the 1947 Act).
As at 6 November 1987 (the relevant sections of the 1947 Act came into operation on 30 October 1987) there were 669,293 voters enrolled in the Metropolitan Area and 240,081 voters enrolled in the remainder of the State. This represented 74 per cent and 26 per cent respectively of the total of enrolled voters. Having regard to the number of members constituting the Assembly, this meant that 74 per cent of the voters would elect 60 per cent of the members and 26 per cent of the voters would elect 40 per cent of the members.
In dividing the Metropolitan Area and the remainder of the State into districts, the Electoral Distribution Commissioners are required to ensure that the number of voters in any district must not be more than 15 per cent greater or more than 15 per cent less than the quotient determined by dividing the number of electors in the area by the number of districts (s 6(2) of the 1947 Act).
From this the plaintiffs argue that, assuming a perfect division into electorates within each area, a metropolitan vote is worth 53 per cent of a non-metropolitan vote.
As to the Legislative Council, mention has been made earlier in this judgment of the fact that it comprises thirty-four members from six regions (s 5 of the 1899 Act). Three regions consist of districts that form the Metropolitan Area. The fourth, the Mining and Pastoral Region, consists of districts "that are remote from the capital and where the land use is primarily for mining and pastoral purposes". The fifth, the Agricultural Region, consists of districts "generally south, or south and west, of and adjacent to the Mining and Pastoral Region". The remaining region is the South West Region. Each region must consist of "complete and contiguous districts" (s 9 of the 1947 Act).
At the time of the 1987 legislation, the effect was that 74 per cent of metropolitan electors would choose 50 per cent of the Legislative Council members while 26 per cent of non-metropolitan electors would choose 50 per cent of the members.
The plaintiffs argue that because of differences in the number of voters in and outside the Metropolitan Area, and having regard to the figures current at the time of the 1989 and 1993 State general elections, the ratios between the highest number of voters in one district of the Legislative Assembly and the lowest in another were 2.52:1 and 2.91:1 respectively. In other words, the numbers of voters in one district were between 252 per cent and 291 per cent of those in another. Likewise, they say, the extremes of ratios of the Legislative Council voters were 3.31:1 and 3.42:1 so that the numbers of voters in one region were between 331 per cent and 342 per cent of those in another. The plaintiffs give colour to their argument by saying that, for example, at the 1993 election the District of Wanneroo contained 291 per cent of the number of voters in the District of Ashburton and that, in terms of current electoral enrolments, the figure is 414 per cent.
The defendant points to the vast area of the State, most of which is thinly populated, and to the difficulties of distance and communication which have the potential to impair effective representation for those who live in the more isolated parts of the State. It argues that to require electorates to be of equal size would result in electorates which were unmanageably large and difficult for the member to service. It argues that these considerations justify the impugned legislative scheme. Its written submissions point to the long history in Western Australia of differentiating between rural and urban electorates. But historical considerations are not conclusive [229] .
The aim of facilitating the representation of those who live in the thinly populated and remote areas of the State is clearly a legitimate one. As McLachlin CJ said in Dixon v Attorney-General (British Columbia) [230] :
[O]nly those deviations should be admitted which can be justified on the ground that they contribute to better government of the populace as a whole, giving due weight to regional issues within the populace and geographic factors within the territory governed. 1. (1989) 59 DLR (4th) 247 at 267.
But while the aim is legitimate, it cannot be said that the legislative means chosen are proportionate to the aim. The impugned legislative scheme arbitrarily distinguishes between metropolitan and non-metropolitan areas. It does not tailor electoral divisions to take account of the difficulties faced by the voters in particularly remote areas [231] . The strength of the plaintiffs' argument lies in the fact that the system mandated by the challenged legislative provisions inevitably produces a serious malapportionment between metropolitan and non-metropolitan electors. That malapportionment results in large part from the absence of flexibility in the system which would allow for a movement in population from rural areas to urban centres and otherwise ensure a more fair representation of urban voters. The vice lies in the absence of flexibility, rather than the need to have regard to particular problems of representation affecting non-metropolitan electors. It was the differential treatment of areas on a generalised basis, rather than by reference to the needs of particular electoral districts, that led to the invalidation of the British Columbia legislation in Dixon v Attorney-General (British Columbia) [232] . It is true that disproportion in the value of a vote may arise in other ways, for instance through the definition of electoral divisions. And this may be deliberate, as in the case of a gerrymander. But these are not issues with which the present proceedings are concerned. It is a particular aspect of representative democracy that the Court is called upon to consider.
It follows that s 6 of the 1899 Act and ss 2A, 6 and 9 of the 1947 Act are together at odds with the principle of representative democracy to be found in the Western Australian Constitution.
This is not the view of the majority of the Court. In the light of their Honours' reasons the answers they propose to the questions asked are inevitable. If the majority view were otherwise, questions would arise as to the orders the Court should make: whether it should answer the questions in the form asked, whether orders of a declaratory nature should be made, whether any orders should be suspended for a time [233] .
The view I take of representative democracy as found in the Western Australian Constitution is that it is not a fixed concept but rather is responsive to the time and circumstances in which it falls for consideration. Therefore, rather than answer the questions as asked, it is appropriate to say no more than that if an election were now held for the Legislative Assembly or Legislative Council in accordance with the impugned statutory provisions, the members would not be chosen by the people in accordance with the constitutional dictates of representative democracy.
The facts, the relevant legislative provisions, their legislative history and the questions in the case stated are set out in the judgment of Toohey J. They need not be repeated. I am in general agreement with his Honour's reasons and shall set out my own reasons only in those areas where my approach differs from his. Like his Honour, I think the preferable course is not to answer the questions in the case stated. However, my approach differs from his Honour's to the extent that I would dispose of the case stated by saying that, if elections were now held for the Legislative Council or for the Legislative Assembly of Western Australia with members being returned for the electoral regions or electoral districts provided for, respectively, by s 6 of the Constitution Acts Amendment Act 1899 WA (the 1899 Act) and ss 2A(2), 6 and 9 of the Electoral Distribution Act 1947 WA (the 1947 Act), neither the members of the Council nor those of the Assembly would be "chosen by the people" within the meaning of those words in s 73(2)(c) of the Constitution Act 1889 WA (the 1889 Act).
I agree with Toohey J that it is the 1889 Act and not the Australian Constitution that is decisive in this case. I am of the view that, when proper regard is had to the system of representative democracy which inheres in the text and structure of the Australian Constitution, s 106 of the Constitution operates to require that the States, as constituent bodies of the federation, be and remain essentially democratic. But that requirement stops considerably short of the proposition advanced by the plaintiffs, namely, that the Constitution, either by reason of s 15 or by reason of s 106, requires that the Parliaments of the States be elected pursuant to voting systems which, within practicable limits, ensure equality of voting power across electorates. Subject to what I have earlier said with respect to s 106, I adopt the reasons of Toohey J for concluding that neither s 15 nor s 106 of the Constitution has that effect.
Although it is s 73(2)(c) of the 1889 Act that is decisive in this case, it is convenient to begin with an analysis of the expression "directly chosen by the people" in ss 7 and 24 of the Australian Constitution, from which the expression "chosen directly by the people" in s 73(2)(c) appears to have been taken. The language is identical, except for the placement of the adverb, and the expressions are both to be found in constitutional documents. Unless there is some difference in context, ordinary principles of construction require that the expressions be construed consistently.
An argument that the expression "directly chosen by the people" in s 24 of the Constitution requires that single member electorates have equal numbers of people or equal numbers of voters was rejected by the majority in Attorney-General (Cth); Ex rel McKinlay v The Commonwealth [234] (Barwick CJ, McTiernan, Gibbs, Stephen, Mason, Jacobs JJ, Murphy J dissenting). Different approaches were taken by the different justices who constituted the majority in that case. Barwick CJ and Gibbs J each emphasised that various constitutional provisions, including ss 7 and 24, either assumed or required inequality, that at the time of federation the franchise was not uniform throughout the colonies and that the history and origins of the Australian Constitution differed significantly from those of the Constitution of the United States of America [235] , Art I, s 2 of which also contains the phrase "chosen by the People" and has been construed as requiring that, as nearly as practicable, one person's vote should be equal to another's [236] . Barwick CJ concluded that "there [was] no constitutional guarantee of a universal franchise of equal voting strength or value, nor of practical equality in electoral divisions either in numbers of people or of electors" [237] . Gibbs J concluded to similar effect, observing that "the framers of the Constitution trusted the Parliament to legislate with respect to [the democratic process] , no doubt remembering that in England, from which our system of representative government is derived, democracy did not need the support of a written constitution" [238] .
In McKinlay , Mason J also referred to constitutional provisions which indicated that inequality was tolerated by the Constitution and to the different history of the Constitution of the United States [239] and concluded that s 24 permitted "some, or even a marked, variation" [240] in electorates. Even so, his Honour thought it was "perhaps conceivable that variations in the numbers of electors or people in single member electorates could become so grossly disproportionate as to raise a question whether an election held on boundaries so drawn would produce a House of Representatives composed of members directly chosen by the people of the Commonwealth" [241] .
It would seem that Stephen J also accepted that there might be circumstances in which disparity between electorates might offend representative democracy which, in his Honour's view, was to be discerned in the expression "directly chosen by the people". Thus, his Honour observed that there are "finite limits" to what is comprehended by representative democracy and that "in a particular instance there may be absent some quality which is regarded as so essential to representative democracy as to place that instance outside those limits altogether" [242] . And, a little later, his Honour added that "something approaching numerical equality of electors within electorates is an important factor in the attainment of what many will regard as representative democracy in its purest form", even though, in its "absolute form" it is not "necessarily imported into the Constitution by the selection of representative democracy as the chosen mode of government for the nation" [243] .
In a joint judgment, McTiernan and Jacobs JJ rejected the proposition that s 24 required "absolute or as nearly as practicable absolute equality of numbers in every constituted electoral district in a State" whilst acknowledging that the "notion of equality [was] present" in the expression "chosen by the people" but that it was a "matter of degree" [244] . Their Honours considered that the meaning and application of s 24 were not frozen by reference to the circumstances existing at federation, saying [245] :
The words "chosen by the people of the Commonwealth" fall to be applied to different circumstances at different times and at any particular time the facts and circumstances may show that some or all members are not, or would not in the event of an election, be chosen by the people within the meaning of these words in s 24. At some point choice by electors could cease to be able to be described as a choice by the people of the Commonwealth. It is a question of degree. It cannot be determined in the abstract. It depends in part upon the common understanding of the time on those who must be eligible to vote before a member can be described as chosen by the people of the Commonwealth. For instance, the long established universal adult suffrage may now be recognised as a fact and as a result it is doubtful whether, subject to the particular provision in s [246] [of the Constitution], anything less than this could now be described as a choice by the people. 1. McKinlay (1975) 135 CLR 1 at 36-37 2. McKinlay (1975) 135 CLR 1 at 36 3. Section 30 of the Constitution is as follows: "Until the Parliament otherwise provides, the qualification of electors of members of the House of Representatives shall be in each State that which is prescribed by the law of the State as the qualification of electors of the more numerous House of Parliament of the State; but in the choosing of members each elector shall vote only once."
One other matter should be noted with respect to McKinlay . Save to the extent that Stephen J was of the view that "representative democracy , direct popular election, and the national character of the lower House, may be discerned in the opening words of s 24" [247] , none of the justices who constituted the majority gave precise content to the words "chosen by the people". Perhaps that is because, as McTiernan and Jacobs JJ pointed out, they do not permit of exact content [248] . Whether or not this be so, the fact that no precise content was given and the different approaches in the judgments require that McKinlay be accepted only as authority for the proposition that s 24 does not require absolute or as near as practicable equality between electorates. More significantly for present purposes, it is not authority for the proposition that the Constitution provides general authority for continuing significant malapportionment between electorates. Rather and save for Barwick CJ and Gibbs J, the justices who constituted the majority in that case appear to have accepted that inequalities might be of such magnitude that, at some point, members of the House of Representatives might not be "chosen by the people" or, in the case of Stephen J, the requirements of representative democracy might not be satisfied [249] .
As already indicated, reference was made in McKinlay to a number of matters which indicate that the Constitution tolerates some disparity in voting strength. There is of course the consideration that, subject to the requirement in ss [250] and [251] , the Constitution expressly allowed for State laws to determine the qualification of electors until the Parliament of the Commonwealth should otherwise provide, and did so in a context where the franchise was far from uniform [252] and full adult suffrage existed only in South Australia [253] . There is also the consideration that, despite population differences between the States, s 7 requires an equal number of senators for each of the Original States [254] while s 24 directs that, notwithstanding its requirement that "[t]he number of members chosen in the several States [for the House of Representatives] shall be in proportion to the respective numbers of their people", five members at least are to be chosen in each Original State [255] .
There is one other feature of s 24 which was the subject of remarks in McKinlay , to which I have yet to refer. As with s 7, s 24 makes no reference to electors, elections or to persons being elected, but to members being "chosen by the people" [256] . This is of some significance in a constitutional context where there are separate provisions with respect to elections [257] , the franchise [258] and the numbers of senators and members of the House of Representatives to be chosen in each State [259] . In that context, particularly when regard is had to the indications that the Constitution tolerates some inequality, the expression "chosen by the people" must be seen as mandating a democratic electoral system and not as requiring a particular electoral system or that it have some particular feature. This notwithstanding, it is clear that there is a requirement in ss 7 and 24 that senators and members of the House of Representatives be "chosen by the people". And that requirement is not satisfied merely by the holding of elections. For example, the Parliament could not legislate pursuant to s 34 of the Constitution to make membership of a particular political party the qualification for election to the House of Representatives. Such a law would so deprive the electorate of choice that persons elected pursuant to it could not be described as "chosen by the people".
It follows from what has been said that there may be some feature of the electoral system or, as Stephen J acknowledged in McKinlay , "there may be absent some quality which is regarded as so essential to representative democracy" [260] that it cannot be said that persons elected pursuant to it are "chosen by the people". The problem is to identify the process by which it may be determined whether or not that is so.
As already indicated, the words "chosen by the people" mandate a democratic electoral system. In that sense, they, like Art I, s 2 of the Constitution of the United States, are a "guarantee of democracy" [261] . It has long been accepted that the Australian Constitution is "broad and general in its terms, intended to apply to the varying conditions which the development of our community must involve" [262] . The words "chosen by the people" are as broad and general as any in the Constitution and, as with other words which are necessarily general, they are to be approached on the basis that, although their essential meaning is unchanged, "their full import and true meaning can often only be appreciated when considered, as the years go on, in relation to the vicissitudes of fact which from time to time emerge" [263] . They must also be interpreted bearing in mind that democracy was not a perfectly developed concept at the time of federation and, perhaps, is not yet so. These considerations necessitate that the content and application of the words "chosen by the people" be determined in the light of developments in democratic standards and not by reference to circumstances as they existed at federation, with the consequence, as McTiernan and Jacobs JJ acknowledged in McKinlay [264] , that what was permitted by s 24 at one time may not be permitted at another.
Notwithstanding the limited nature of the franchise in 1901 [265] , present circumstances would not, in my view, permit senators and members of the House of Representatives to be described as "chosen by the people" within the meaning of those words in ss 7 and 24 of the Constitution if the franchise were to be denied to women or to members of a racial minority or to be made subject to a property or educational qualification. Similarly, the fact that the Constitution countenanced disparity in voting value at the time of federation provides no basis for a conclusion that significant disparity in the numbers of electors in single member electorates does not now offend s 24.
It does not follow from what has been said that s 24 requires complete or even practical equality of electorate size. A distinction which is reasonably capable of being viewed as an appropriate and adapted means of taking account of geographic boundaries, community or minority interests [266] or some other matter which bears on effective parliamentary representation, such as the dispersed nature of the population in remote areas, would not, in my view, prevent it being said that members of the House of Representatives were "chosen by the people". However, subject to that consideration and the requirements of the Constitution which necessitate or which may necessitate inequality by reason of population differences between the States [267] , persons elected under a system involving significant disparity in voting value, could not, in my view, now be described as "chosen by the people". To so describe them would be contrary to current democratic standards reflected in the electoral laws of the Commonwealth and, putting Western Australia to one side, the electoral laws of the States referred to in the judgment of Toohey J.
As already indicated, unless there is some relevantly different context, the words "chosen by the people" in s 73(2)(c) of the 1889 Act must be interpreted consistently with those same words in ss 7 and 24 of the Constitution. There is one difference: when s 73(2)(c) was inserted in the 1889 Act [268] there was no constitutional provision which necessitated or, depending on population differences, might necessitate disparity in voting value, as is the case with ss 7 and 24 which, respectively, require that there be an equal number of senators for each of the Original States and that at least five members of the House of Representatives be chosen in each Original State. This means that there is no basis for any constitutional limitation in relation to voting equality of the kind necessary in relation to ss 7 and 24 of the Constitution. That aside, the words "chosen by the people" in s 73(2)(c) of the 1889 Act must, in my view, be applied in the same way as they are in ss 7 and 24 of the Constitution. More particularly, they must be viewed as constituting a guarantee of democracy entrenched in the 1889 Act unless and until amended in accordance with s 73(2). And they must be applied in the light of contemporary circumstances, having regard to developments which have taken place in democratic standards.
In applying the words "chosen by the people", it is relevant to have regard to the fact that, in 1978 [269] , when those words were inserted in the 1889 Act, the 1947 Act allowed for significant disparity in voting value both for Legislative Council and Legislative Assembly elections [270] . The extent of that disparity and the fact that it was countenanced until comparatively recent times militates, at least for the moment, against an approach to s 73(2)(c) which would require precisely the same standard as that now required by s 24 of the Constitution. Rather, that history provides a basis, at least for the moment, for a somewhat more generous margin of variation in voting value. But as with s 24 of the Constitution, the countenancing of voting inequality at one time does not, of itself, provide a basis for countenancing inequality at a later time. Ultimately, it is a question of what is required in the light of current democratic standards, including those which so recently applied in Western Australia.
In my view, the malapportionment which is detailed in the judgment of Toohey J is so great as to be distinctly at odds with democratic standards revealed in the electoral laws of the Commonwealth and the other Australian States referred to in his Honour's judgment. Moreover, the distinction between metropolitan and non-metropolitan areas is, as his Honour points out, arbitrary and inflexible and, that being so, it cannot be justified on the basis that it is reasonably capable of being seen as appropriate and adapted to the dispersed nature of the population in the remote regions of Western Australia or to any other matter or circumstance which might bear on effective parliamentary representation.
It was not argued that s 6 of the 1899 Act and ss 2A(2), (6) and (9) of the 1947 Act were invalid when enacted in 1987 or, indeed, at any particular time thereafter. Rather, the plaintiffs' case was that, at some unspecified time, the provisions became invalid and remain so, in the same way, presumably, as laws which are validly enacted under the defence power may cease to be laws with respect to defence and become void when hostilities have ceased [271] . I am not convinced that that is the proper approach to be taken with respect to electoral laws which are put into operation as and when elections have to be held.
The issue that arises with respect to s 6 of the 1899 Act and ss 2A(2), (6) and (9) of the 1947 Act is, in essence, whether they would have valid operation if elections were now held. Because that is the issue, the preferable course, in my view, is to answer the question in the case stated by saying, as earlier indicated, that if elections were now held, neither the members of the Legislative Council nor those of the Legislative Assembly would be "chosen by the people" within the meaning of those words in s 73(2)(c) of the 1889 Act. This is a minority view and it is, thus, unnecessary to consider what, if any, further relief is required by the answer I propose.
The question in this case stated is whether, in an election for the Legislative Council or Legislative Assembly for the State of Western Australia, the number of persons eligible to vote for or required to elect a member in one electorate should be approximately the same as the number of persons eligible to vote for or required to elect a member in another electorate.
The plaintiffs contend that, under the Commonwealth of Australia Constitution Act 1900 (the Constitution) and the Constitution Act 1889 WA (the Western Australian Constitution), every legally capable adult person is entitled to vote in such an election and that each vote, so far as is reasonably practicable, should be equal in value. The plaintiffs contend that both Constitutions contain an implied principle of representative democracy and that that principle entails the right of each individual elector to participate in the electoral process on an equal basis. They contend that s 6 of the Constitution Acts Amendment Act 1899 WA (the 1899 Act) and ss 2A(2), 6 and 9 of the Electoral Distribution Act 1947 WA (the 1947 Act) permit breaches of this right and that they are invalid.
The Acts Amendment (Electoral Reform) Act 1987 WA (the Amending Act) came into operation on 30 October 1987. It amended provisions of the 1899 Act and the 1947 Act which regulate the constitution of the Legislative Council and Legislative Assembly and which provide for the division of the State into electoral regions. Section 6 of the 1899 Act provides that the State shall be divided into six electoral regions under the 1947 Act. Section 2A of the latter Act requires the State to be divided into districts and regions in accordance with the Act. Section 9 of that Act directs Electoral Distribution Commissioners to divide the State into six regions. Three regions are to be known respectively as the North Metropolitan Region, the South Metropolitan Region and the East Metropolitan Region and are to "consist of complete and contiguous districts that together form the Metropolitan Area" [272] . The fourth region is to be known as the Mining and Pastoral Region and is to consist "of complete and contiguous districts that are remote from the capital and where the land use is primarily for mining and pastoral purposes". The fifth region is to be known as the Agricultural Region and is to consist of complete and contiguous districts "that together form an area that is generally south, or south and west, of and adjacent to the Mining and Pastoral Region". The remaining region is to be known as the South West Region and is to consist of complete and contiguous districts.
Under s 6 of the 1899 Act, the electoral regions known as the North Metropolitan Region and the South West Region are each to return seven members, and the electoral regions known as the South Metropolitan Region, the East Metropolitan Region, the Agricultural Region and the Mining and Pastoral Region are each to return five members, to serve in the Legislative Council. As a result, the Legislative Council consists of thirty-four members elected from six regions. Three regions are located in the Metropolitan Area and three outside that area. The Metropolitan regions return seventeen members; so do the three regions outside the Metropolitan Area.
Section 19 of the 1899 Act also provides that, following the first dissolution of the Legislative Assembly or the expiry of its term after 31 December 1982, the State is to be divided in fifty-seven electoral districts under the provisions of the 1947 Act with each district to return one member to serve in the Legislative Assembly. Section 6(1) of the 1947 Act directs the Commissioners to divide the Metropolitan Area into thirty-four districts and to divide the remainder of the State into twenty-three districts. Section 6(2) directs the Commissioners to make the division of an area into districts "in accordance with the principle that the number of enrolled electors comprised in any district in the area must not be more than 15 per cent greater, or more than 15 per cent less, than the quotient obtained by dividing the total number of enrolled electors in the area by the number of districts into which the area is to be divided". As a result of these enactments and the distributions made under them, the Legislative Assembly consists of fifty-seven members, each elected for a single district, thirty-four of whom are returned for districts in the Metropolitan Area and twenty-three for the rest of the State.
As at 6 November 1987, 669,293 people, constituting 74 per cent of the total voters in Western Australia, were enrolled in the Metropolitan Area and 240,081 voters, constituting 26 per cent of that total, were enrolled for the remainder of the State. Consequently, as at that date, 74 per cent of voters in the Metropolitan Area elected 60 per cent of the members of the Legislative Assembly while 26 per cent of voters outside the Metropolitan Area elected 40 per cent of its members. As the plaintiffs point out, the boundary of the Metropolitan Area is fixed for electoral purposes as at 1 January 1987 irrespective of the number of people who reside in it.
As far as the Legislative Council was concerned, the result of the legislation was that metropolitan electors, totalling 74 per cent of the State's voters, were to elect the same number of Legislative Council members as the non-metropolitan electors who comprised 26 per cent of the State's voters.
The 15 per cent tolerance from the average quotient together with the inevitable changes in population produces large divergences in the number of voters enrolled for districts in the metropolitan and non-metropolitan areas of the State. At the date of the 1993 election, the quotient for the Metropolitan Area was 21,988 with the 15 per cent tolerance giving a range of 18,690 to 25,286 voters for each district. In Belmont, the smallest district, 20,064 voters were enrolled while in Wanneroo, the largest district, 26,580 voters were enrolled. In the non-metropolitan areas of the State, the quotient was 11,702 with the 15 per cent tolerance giving a range of 9,947 to 13,457 voters for each district. In Ashburton, the smallest district, 9,135 voters were enrolled for the 1993 election while in Murray, the largest district, 15,646 voters were enrolled.
Thus, at the 1993 State election, the electoral district of Wanneroo contained 291 per cent more voters than the electoral district of Asburton. Moreover, because the State was divided into the Metropolitan Area and the rest of the State, disparities in the number of voters in electoral districts existed even in respect of districts which were geographically close to each other. In the district of Peel, for example, 25,547 voters were enrolled while only 11,045 voters were enrolled in the nearby district of Mandurah.
The elections for the Legislative Council disclosed even greater discrepancies between the numerical size of the regions that elect members. Thus the quotient for the North Metropolitan Region was 34,161 votes and for the Mining and Pastoral Region 9,097. In other words, the ratio between the number of votes required to elect a member in these two regions was 3.76 to 1.
It is beyond question that the distribution of voters in the electoral districts and regions of Western Australia, for both the Legislative Council and the Legislative Assembly, results in the vote of some voters, particularly non-metropolitan voters, having a greater value than that of others. If the principle of representative democracy is a principle of the Constitution or the Western Australian Constitution, and if representative democracy under either Constitution requires that the number of voters in electoral districts or regions should be equal so far as is reasonably practicable, the provisions of the 1899 Act and the 1947 Act are in breach of those Constitutions. The scheme set up by the two Acts arbitrarily distinguishes between metropolitan and non-metropolitan voters. On no rational basis can the special needs of electors in areas outside the non-metropolitan areas justify such large disparities as exist between particular electoral districts and regions. It is unnecessary, however, for me to develop these points because, contrary to the plaintiffs' submissions, I am of the opinion that neither the Constitution nor the Western Australian Constitution contains any requirement that the number of voters in electoral districts should be equal or equal so far as is reasonably practicable.
In Attorney-General (Cth); Ex rel McKinlay v The Commonwealth [273] , this Court (Barwick CJ, McTiernan, Gibbs, Stephen, Mason and Jacobs JJ, Murphy J dissenting) held that the Constitution did not require the number of people or the number of electors in electoral divisions for elections to the federal House of Representatives to be equal. In McKinlay , the plaintiff argued that equality of voters in electoral divisions was required by s 24 of the Constitution which provides, so far as material, that "[t]he House of Representatives shall be composed of members directly chosen by the people of the Commonwealth". The plaintiff in that case did not rely upon an implied principle of representative democracy to support his claim, no doubt for the reason that at that time it had not occurred to him or his advisers that the Constitution contained such a principle.
The plaintiffs in this case contend that the decision in McKinlay is no longer an authority on the meaning of s 24 of the Constitution because it has been overtaken by those decisions of this Court which hold that the principle of representative democracy is inherent in the Constitution independently of the terms of any particular provision of that Constitution [274] . The plaintiffs contend that, once this principle is accepted, there is no reason for distinguishing those decisions [275] of the United States Supreme Court that hold that the United States Constitution contains a principle of equality of voting power. They argue that the decisions of the United States Supreme Court on equality of electoral districts are directly relevant to Australia because the Constitution, like the United States Constitution, has adopted a structure of representative government. The plaintiffs also rely on the similarity between the terms of s 24 of the Constitution and Art I, s 2 of the United States Constitution which enacts that the House of Representatives "shall be composed of Members chosen every second Year by the People of the several States". They point out that the words "chosen by the people" are common to both Constitutions.
In Wesberry v Sanders [276] , Black J, speaking for the United States Supreme Court, said:
We hold that, construed in its historical context, the command of Art 1, s 2, that Representatives be chosen "by the People of the several States" means that as nearly as is practicable one man's vote in a congressional election is to be worth as much as another's To say that a vote is worth more in one district than in another would not only run counter to our fundamental ideas of democratic government, it would cast aside the principle of a House of Representatives elected "by the People"
While it may not be possible to draw congressional districts with mathematical precision, that is no excuse for ignoring our Constitution's plain objective of making equal representation for equal numbers of people the fundamental goal for the House of Representatives. (Footnotes omitted.)
The equal representation for equal numbers principle has been rigorously applied by the Supreme Court. In White v Weiser [277] , the Court invalidated a plan which allowed a difference of 4.2 per cent between the most populous and least populous congressional district in Texas. In Karcher v Daggett [278] , the Court was even more rigorous. It invalidated a plan in which the difference between the largest district and the smallest district in New Jersey was only 0.7 per cent.
The plaintiffs contend that the statement of Warren CJ in Reynolds v Sims [279] is equally applicable to our Constitution. His Honour said:
The right to vote freely for the candidate of one's choice is of the essence of a democratic society, and any restrictions on that right strike at the heart of representative government. And the right of suffrage can be denied by a debasement or dilution of the weight of a citizen's vote just as effectively as by wholly prohibiting the free exercise of the franchise. 1. (1964) 377 US 533 at 555.
Legislators represent people, not trees or acres. Legislators are elected by voters, not farms or cities or economic interests. As long as ours is a representative form of government, and our legislatures are those instruments of government elected directly by and directly representative of the people, the right to elect legislators in a free and unimpaired fashion is a bedrock of our political system. It could hardly be gainsaid that a constitutional claim had been asserted by an allegation that certain otherwise qualified voters had been entirely prohibited from voting for members of their State legislature. And, if a State should provide that the votes of citizens in one part of the State should be given two times, or five times, or 10 times the weight of votes of citizens in another part of the State, it could hardly be contended that the right to vote of those residing in the disfavoured areas had not been effectively diluted. 1. Reynolds v Sims (1964) 377 US 533 at 562
The plaintiffs also rely on decisions [281] of the United States Supreme Court that hold that the equal protection clause in s 1 of the Fourteenth Amendment to the United States Constitution requires that, whenever the members of a legislature are elected from districts within a State or city or local government area, each district, so far as practicable, should have an equal number of voters who elect a proportionately equal number of members. The plaintiffs assert that the notion of equal protection of the laws which is expressly enacted in the United States Constitution is implied in the Constitution.
In my opinion, the United States decisions do not govern the construction of the Constitution. The decisions on Art I, s 2 were heavily influenced by United States history, particularly the intentions of the framers of the United States Constitution [282] . The Australian history is different. The equal protection clause which is the basis of the other decisions of the Supreme Court has no counterpart in the Constitution. Moreover, the applicability of Baker v Carr [283] , Reynolds v Sims and Wesberry v Sanders to s 24 of our Constitution was considered and rejected by a majority of the Court in McKinlay [284] .
The Constitution contains no express requirement that the number of electors in electoral divisions for federal or State elections should be numerically equal or numerically equal so far as is practicable, nor do the terms of the Constitution or its structure nor the history of elections at the time of federation provide any foundation for such a conclusion. On the contrary, various provisions of the Constitution convincingly point to the conclusion that neither equality of voting power nor equal numbers of electors in divisions electing members to the houses of the federal or State parliaments is a constitutional requirement. If the principle of representative government or representative democracy is part of the Constitution independently of the terms of ss 1, 7, 24, 30 and 41 of that Constitution, that principle, because it arises from implication, must give way to the inferences to be drawn from other provisions of the Constitution. Those provisions show that neither equality of voting power nor equal numbers of electors in divisions within a State is a constitutional requirement for either federal or State elections.
The Constitution is contained in a statute of the United Kingdom Parliament. In the late twentieth century, it may not be palatable to many persons to think that the powers, authorities, immunities and obligations of the federal and State parliaments of Australia derive their legal authority from a statute enacted by the Imperial Parliament, but the enactment of that statute containing the terms of the Constitution is the instrument by which the Australian people have consented to be governed. Since the passing of the Australia Act UK in 1986, notwithstanding some considerable theoretical difficulties [285] , the political and legal sovereignty of Australia now resides in the people of Australia [286] . But the only authority that the people have given to the parliaments of the nation is to enact laws in accordance with the terms of the Constitution.
The Constitution contains no injunction as to how it is to be interpreted. Any theory of constitutional interpretation must be a matter of conviction based on some theory external to the Constitution itself. But since the people have agreed to be governed by a constitution enacted by a British statute, it is surely right to conclude that its meaning must be determined by the ordinary techniques of statutory interpretation and by no other means [287] . It must therefore be interpreted by late twentieth century Australians according to the ordinary and natural meaning of its text, read in the light of its history, with such necessary implications as derive from its structure. Nevertheless, as Higgins J pointed out [288] in the first decade of this century:
[A]lthough we are to interpret the words of the Constitution on the same principles of interpretation as we apply to any ordinary law, these very principles of interpretation compel us to take into account the nature and scope of the Act that we are interpreting - to remember that it is a Constitution, a mechanism under which laws are to be made, and not a mere Act which declares what the law is to be. (Emphasis in original.) In similar vein in Jumbunna Coal Mine NL v Victorian Coal Miners' Association [289] , O'Connor J stated a principle of constitutional interpretation which has frequently been approved [290] and applied in this Court. His Honour said [291] : [I]t must always be remembered that we are interpreting a Constitution broad and general in its terms, intended to apply to the varying conditions which the development of our community must involve.
For that reason, where the question is whether the Constitution has used an expression in the wider or in the narrower sense, the Court should, in my opinion, always lean to the broader interpretation unless there is something in the context or in the rest of the Constitution to indicate that the narrower interpretation will best carry out its object and purpose.
The ordinary principles of statutory interpretation require that the text be the starting point of any interpretation of the Constitution. Part of the ordinary and natural meaning of the text is any implication which is "manifested according to the accepted principles of interpretation" [292] . Implications derived from the structure of the Constitution are also part of the Constitution's meaning but such implications may be drawn only when they are "logically or practically necessary for the preservation of the integrity of that structure" [293] . Thus, because the Constitution has prescribed a system for elections to the Houses of the federal Parliament, no Australian government can pass laws which would undermine the efficacy of that system [294] . However, I cannot accept, as Deane and Toohey JJ held in Nationwide News Pty Ltd v Wills [295] , that a constitutional implication can arise from a particular doctrine that "underlies the Constitution". Underlying or overarching doctrines may explain or illuminate the meaning of the text or structure of the Constitution but such doctrines are not independent sources of the powers, authorities, immunities and obligations conferred by the Constitution. Top-down reasoning [296] is not a legitimate method of interpreting the Constitution. As I pointed out in Theophanous v Herald & Weekly Times Ltd [297] , after the decision of this Court in the Engineers' Case [298] , the Court had consistently held, prior to Nationwide News and Australian Capital Television Pty Ltd v The Commonwealth [299] , that it is not legitimate to construe the Constitution by reference to political principles or theories that are not anchored in the text of the Constitution or are not necessary implications from its structure. I pointed out that the Engineers' Case had made it plain that the Constitution was not to be interpreted by using such theories to control or modify the meaning of the Constitution unless those theories could be deduced from the terms or structure of the Constitution itself. It is the text and the implications to be drawn from the text and structure that contain the meaning of the Constitution. With all due respect to the judges of this Court who have held that there is a free-standing principle of representative democracy in the Constitution, their conclusion necessarily involves a rejection of the principles of interpretation laid down in the Engineers' Case [300] although perhaps not the philosophy that lies behind that decision [301] .
One great difficulty with the implication of representative democracy that has recently been discovered in the Constitution is the nature of the implication. In a constitutional instrument, such as the Constitution, an implication will ordinarily confer a power or a right on a legislature or person or restrict the exercise of a power or a right, or in some cases impose an obligation, on a legislature or person. Prior to the series of cases which commence with Nationwide News in 1992, implications drawn from the Constitution were seen as either conferring power on the Commonwealth or as restraining the exercise of power by the Commonwealth or the States. Thus, certain powers were seen as implied in the Commonwealth from "the very nature and existence of the Commonwealth as a political institution" [302] and certain restraints on the Commonwealth [303] or the States [304] were seen as implied from the structure of the Constitution. But the implication of a principle of representative democracy is a different sort of implication.
After the decision in Australian Capital Television and prior to the decision in Theophanous , the precise implication might have been thought to be that no government, State or federal, can legislate in a way that undermines or interferes with the system of representative government so far as the Constitution sets it up. In that respect, the implied principle of representative democracy would be a shorthand but inexact expression for an implication derived from the need to maintain the representative system of federal government that the Constitution sets up. It would be similar to the implied principle of the separation of powers which is a shorthand expression for saying that neither the legislature nor the executive can exercise judicial power nor can the legislature impose non-judicial functions on the judiciary. There was, however, difficulty in characterising the implication of representative democracy in this way because in Australian Capital Television a majority of justices took the view that the principle impliedly restricted the grant of affirmative powers in s 51 of the Constitution if those powers interfered with the State and local government systems of government. This was because the three tiers of government were seen to be closely tied together and it was regarded as unrealistic to attempt to distinguish the federal system from the other systems of government.
Since the decisions of the Court in Theophanous and Stephens v West Australian Newspapers Ltd [305] , however, it is plain that the implication cannot be likened to the implication of the separation of powers. Indeed in those two cases, the implication seems to have taken on another dimension, if it has not changed dramatically; for in those cases, the Court has held that the implied principle is not only a restraint on federal and State legislative power but it changes common law rights and duties. That would seem to mean that the Constitution directly changes the rights and duties of individuals. However, the majority of justices in Theophanous [306] expressly said that:
Whether the implied freedom could also conceivably constitute a source of positive rights was not a question which arose for decision in those cases and it is unnecessary to decide it in this case. For that reason we shall refer to the freedom of communication as an implication rather than as a guarantee of freedom, notwithstanding the use of the latter expression in some judgments in the two cases.
It is also clear that the implied freedom is one that shapes and controls the common law. At the very least, development in the common law must accord with its content.
The result seems to be that the Constitution contains by implication a principle of representative democracy that is not confined to restricting the powers of the federal or State legislatures, nor does it necessarily confer any rights on individuals. It appears to be a free-standing principle, just as if the Constitution contained a Ch IX with a s 129 which read:
Subject to this Constitution, representative democracy is the law of Australia, notwithstanding any law to the contrary.
That does not mean, of course, that the implied principle of representative democracy trumps all other rights and obligations. Because the principle arises by implication, it must be subject to the express terms of the Constitution and be weighed in appropriate cases against other implications drawn from the text and structure of the Constitution. Nevertheless, when it is seen as the equivalent of the hypothetical s 129, it is plain that it is not an implication contained in the text of the Constitution, nor is it an implication "logically or practically necessary for the preservation of the integrity of [the] structure" of the Constitution [307] . The "implication" has become a premise from which other implications are drawn. Thus, in Theophanous and Stephens the majority held that the preservation of representative democracy required that the common law of defamation should be restricted in its application to federal and State politicians and certain other persons. By that stage the text and structure of the Constitution have receded into the background and it is the concept of representative democracy, not the text or structure of the Constitution, that governs the application of the Constitution in such cases.
The plaintiffs seemed to accept that the principle was a free-standing constitutional provision somewhat similar to the hypothetical s 129 because they argued that, although the connotation of the concept of representative democracy "may be constant, in terms of the purposes, goals and objectives that it entailed for the future of the people of Australia its denotation has been capable of adapting in an evolutionary way to new situations and standards in accordance with the growth of the Australian people". In other words, the content of the "implication" changes with the growth or contraction of the external political theory of representative democracy irrespective of whether the outcome of the growth or contraction is necessary to maintain the system of government set up by the Constitution.
This argument of the plaintiffs may be the logical consequence of the way the majority of the justices of this Court have now formulated the principle, but, to my mind, it only demonstrates that no such principle can be deduced from the Constitution by applying the standard techniques of statutory interpretation to a constitutional instrument.
The authority of the implied principle of representative democracy
Judges as diverse in their approach to the Constitution as Barwick CJ and Murphy J have stated that it is the duty of justices of this Court to apply its text and not the judicial decisions on the text [308] . Judicial exposition of the Constitution does not replace the Constitution or its meaning. As Isaacs J declared in Australian Agricultural Co v Federated Engine-Drivers and Firemen's Association of Australasia [309] :
Our sworn loyalty is to the law itself, and to the organic law of the Constitution first of all. If, then, we find the law to be plainly in conflict with what we or any of our predecessors erroneously thought it to be, we have, as I conceive, no right to choose between giving effect to the law, and maintaining an incorrect interpretation. It is not, in my opinion, better that the Court should be persistently wrong than that it should be ultimately right. But, as Gibbs J pointed out in Queensland v The Commonwealth [310] in a passage to which I referred in Re Tyler; Ex parte Foley [311] , no justice of this Court "is entitled to ignore the decisions and reasoning of his predecessors, and to arrive at his own judgment as though the pages of the law reports were blank". As Gibbs J went on to say, it is only after the most careful consideration of the earlier decision and after giving due weight to all the circumstances that a justice can give effect to his or her own opinion in preference to the earlier decision of the Court. 1. Damjanovic & Sons Pty Ltd v The Commonwealth (1968) 117 CLR 390 at 396 Buck v Bavone (1976) 135 CLR 110 at 137 2. (1913) 17 CLR 261 at 278. 3. (1977) 139 CLR 585 at 599. 4. (1994) 181 CLR 18 at 38.
I regard the reasoning in Nationwide News, Australian Capital Television, Theophanous and Stephens in so far as it invokes an implied principle of representative democracy as fundamentally wrong and as an alteration of the Constitution without the authority of the people under s 128 of the Constitution. Moreover, as much as I admire the noble vision of the justices who have found, contrary to what the overwhelming majority of lawyers had always thought, that the Constitution contains a free-standing principle of representative democracy, the principles of constitutional interpretation compel me to reject their reasoning. It may be that ultimately the representative democracy line of reasoning in Nationwide News and subsequent cases will be so widely followed and applied that, however erroneous one may think that reasoning is, it must be taken to reflect the meaning of the Constitution. But until that time arrives, I conceive that I have no option but to reject the authority of that reasoning.
To decide cases by reference to what the principles of representative democracy currently require is to give this Court a jurisdiction which the Constitution does not contemplate and which the Australian people have never authorised. Interpreting the Constitution is a difficult task at any time. It is not made easier by asking the justices of this Court to determine what representative democracy requires. That is a political question and, unless the Constitution turns it into a constitutional question for the judiciary, it should be left to be answered by the people and their elected representatives acting within the limits of their powers as prescribed by the Constitution.
However, even if, contrary to my view, the Constitution contains a free-standing principle of representative democracy, other provisions of the Constitution show that, whatever the content of representative democracy may be in the last decade of the twentieth century, the Constitution does not require an equal number of electors in electoral divisions. If representative democracy, as presently understood, requires that, so far as is practicable, no group of electors should be able to elect proportionately more elected representatives than a similar group, then our Constitution refuses to give effect to that requirement.
Equality of voting power is not a fundamental feature of the Constitution. On the contrary, inequality of individual voting power is one of its striking features. Thus, under s 128 of the Constitution, an amendment cannot be made to the Constitution unless a majority of electors in a majority of States approve the proposed law to amend the Constitution. Accordingly, the votes of the persons living in one of the less populous States are equivalent to the votes of the persons living in one of the more populous States. If the majority of New South Wales voters approve a constitutional amendment but a majority of Tasmanians reject it, their votes cancel each other out, notwithstanding the enormous population difference between the two States. Thus, the share in the right to amend the Constitution of each Tasmanian voter can be more than twelve times as great as the share of each New South Wales voter.
Further, where an alteration of the Constitution would in any manner affect the provisions of the Constitution in relation to a State, s 128 of the Constitution provides that the alteration shall not be valid unless the majority of electors in the State concerned approved the proposed alteration. In that class of constitutional amendment, the votes of a majority of people in the State concerned are equivalent to the votes of the rest of the nation.
Under the Constitution, therefore, individual Australians do not have an equal share in the sovereignty of Australia. Lord Bryce asserted that, in a country governed by a rigid Constitution which limits the power of the legislature to certain subjects or forbids it to transgress certain fundamental doctrines, the sovereignty of the legislature is necessarily restricted. In that case, ultimate sovereignty resides in the body which made and can amend the Constitution [312] . On that view, the sovereignty of Australia originally resided in the United Kingdom Parliament. Since the Australia Act 1986 UK, however, the sovereignty of the Australian nation has ceased to reside in the Imperial Parliament and has become embedded in the Australian people [313] . Only the people can now change the Constitution. They are the sovereign. But, because their rights to amend the Constitution are not equal, the Australian people do not have equal shares in that sovereignty.
Part II of Ch 1 of the Constitution, which deals with the Senate, is also a powerful indicator that equality of individual voting power is not a feature, let alone a fundamental feature, of the Constitution. Section 7 provides for the equal representation of the States and declares "that equal representation of the several Original States shall be maintained and that no Original State shall have less than six senators". Thus the Senate vote of an elector in Tasmania is ten times more valuable than the Senate vote of an elector in Victoria [314] . At the Sydney session of the Constitutional Convention in 1897, New South Wales suggested that proportional instead of equal representation for a State should be the rule, but the New South Wales proposal attracted only five supporters [315] .
Representation of the Territories in the federal Houses of Parliament also shows that equality of representation for equal numbers is not a requirement of the Constitution. The Constitution does not require the people of the Territories to be represented in the Parliament, but, if it thinks fit, Parliament "may allow the representation of such territory in either House of the Parliament to the extent and on the terms which it thinks fit" [316] . The Territories are not required to be represented in accordance with their population [317] . The words "chosen by the people of the Commonwealth" in s 24 of the Constitution do not include the people of the Territories [318] .
Section 121 of the Constitution is another strong indication that equality of individual voting power is not a general requirement of the Constitution or s 24. It provides that the Parliament may admit new States to the Commonwealth upon "such terms and conditions, including the extent of representation in either House of the Parliament, as it thinks fit". The words of s 24 must give way to this express provision in s 121 just as this Court held in the Territorial Senators Cases [319] that the words of s 7 of the Constitution must give way to the words of s 122, which are similar to s 121, in relation to the representation of the Territories in the Senate. Thus under s 121 as under s 122, in the case of the Territories, the Parliament may prescribe a number of representatives for a new State without reference to the population of the new State or the population of the rest of Australia. The House of Representatives may therefore be composed of members from States and Territories whose share of representation is not proportional to their population. In the face of ss 121 and 122, it is difficult to imply in s 24 a requirement that electoral divisions for elections to the House of Representatives must be equal in numbers.
When the share of individuals in the sovereignty of the nation, in the right to amend the Constitution and in Senate voting is expressly made unequal and when new States and Territories may have unequal representation in the Parliament, it is obvious that equality of individual voting power is not and has not been a fundamental feature of the Constitution. It is therefore unsurprising that the Constitution contains no express indication that each person in a State must have an equal share in electing the total number of members of the House of Representatives for that State.
However, the plaintiffs placed great reliance on the words "directly chosen by the people" in ss 7 and 24 of the Constitution as indicating that the Constitution requires that, within a State, there shall be equal representation for equal numbers. Section 7 provides:
The Senate shall be composed of senators for each State, directly chosen by the people of the State, voting, until the Parliament otherwise provides, as one electorate.
The House of Representatives shall be composed of members directly chosen by the people of the Commonwealth.
The ordinary and natural meaning of those words says very little about how the members are to be chosen. They disclose nothing as to whether voting is to be "first past the post" or preferential voting or whether the electorate is to vote as one electorate or by occupation or by locality. Recourse must be had to history and context to give meaning to the words "directly chosen by the people". For present purposes, it is sufficient to deal with the context and history of those words in s 24.
Part III of Ch 1 of the Constitution, dealing with "The House of Representatives" shows that the context of s 24 is inconsistent with the notion that the number of electors in the electoral divisions of a State must be equal. The provisions of Pt III of Ch 1 demonstrate that the Constitution intended that the manner in which members of the House of Representatives were to be elected was to be a matter for the States unless and until the federal Parliament legislated on the matter. Section 29 of the Constitution provides:
Until the Parliament of the Commonwealth otherwise provides, the Parliament of any State may make laws for determining the divisions in each State for which members of the House of Representatives may be chosen, and the number of members to be chosen for each division. A division shall not be formed out of parts of different States.
In the absence of other provision, each State shall be one electorate.
This section gave each State the power to determine the size of the electorates which were to elect the members elected for that State. It is impossible to construe the general words of s 29 as impliedly requiring a State to have electoral divisions with the same number of electors. There is even less ground for imposing such a restriction on the words "[u]ntil the Parliament otherwise provides" which is the source of the federal Parliament's power to legislate with respect to electoral divisions.
Similarly, s 30 of the Constitution enacted that, until the Parliament otherwise provided, the qualifications of electors for the House of Representatives were to be in each State those which were prescribed by the law of that State as the qualification for the electors of the more numerous House of Parliament of that State. Section 30 also provides that, in the choosing of members, each elector should vote only once. Since the Constitution specifically prohibits an elector voting more than once, it is inconceivable that, if the Constitution intended that the number of electors electing a member for each electoral division should also be equal, it contains no provision to that effect.
The plaintiffs also relied on the requirement in s 24 that:
The number of members chosen in the several States shall be in proportion to the respective numbers of their people. But this requirement of s 24 shows no more than there must be interstate equality; it says nothing about equality of electoral divisions within a State. Moreover, because of the terms of s 121, this requirement of proportionality seems confined to the original States.
As Mason J pointed out in McKinlay [320] , it is significant that:
[T]he second paragraph of s 24 makes explicit provision for the ascertainment of a quota of people for the specific yet limited purpose of ascertaining the number of members to be chosen in each State. Had it been intended to provide that electoral divisions within a State should contain an equal or practically equal number of people or electors it is inconceivable that the quota provision would have been expressed to have such a limited application. 1. (1975) 135 CLR 1 at 62.
The history of electoral distributions prior to federation confirms the view that s 24 of the Constitution does not mandate equality of individual voting power in federal government elections. At federation, equality of electoral populations was not basic to the political systems of England or Australia even though representative government was firmly established in England and in the colonies. Representative government was not perceived as requiring that electoral divisions should be numerically equal. Nor was representative government perceived as requiring that all adult persons should be entitled to vote.
Historically, parliamentary representatives were seen as representing communities and economic interests rather than individual electors [321] . Thus, James Mill thought that governments should be directed by a parliament that, "must have an identity of interest with the community" [322] . During the course of the nineteenth century, however, it came to be seen more clearly that parliamentary representatives should represent individuals rather than communities and economic interests, a view with which Warren CJ agreed when he said in Reynolds v Sims [323] :
Legislatures represent people, not trees or acres. Legislatures are elected by voters, not farms or cities or economic interests. 1. Birch, Representative and Responsible Government: An Essay on the British Constitution (1964), pp 23-24; Maitland, Constitutional History of England (1908), p 363; Reeve and Ware, Electoral Systems: A Comparative and Theoretical Introduction (1992), p 49. 2. An Essay on Government (1820), p 35 quoted in Birch, Representative and Responsible Government: An Essay on the British Constitution (1964), p 46. 3. (1964) 377 US 533 at 562.
In his essay Considerations on Representative Government, John Stuart Mill, who was a believer in the Hare system or proportional representation system of voting [324] , advocated equality of voting power. Equality of electoral districts was also advocated by the radicals under Joseph Chamberlain [325] . Nevertheless, although radical groups continued to advocate for numerical equality of constituencies throughout the nineteenth century, equality of numbers in electoral districts was not a strong issue in 1900. Once the view was accepted that the individual citizen and not the community or some economic interest was the unit of representation, "the extent of the franchise became the main subject of debate" [326] . As Birch says [327] :
The advocacy of equal electoral districts was the important part of the Radical programme in the 1830's but by the turn of the century most of the steam had gone out of this argument. As noted, the redistribution of 1885 had greatly reduced the inequality of constituencies and Liberals were prepared to accept that, up to a point, the principle of equality had to give way to the claims of local interests and administrative convenience. 1. Birch, Representative and Responsible Government: An Essay on the British Constitution (1964), pp 62-63. 2. Chamberlain et al, The Radical Programme [1885], edited with an introduction by D A Hamer (1971), p 21. 3. Birch, Representative and Responsible Government: An Essay on the British Constitution (1964), p 24. 4. Birch, Representative and Responsible Government: An Essay on the British Constitution (1964), p 70.
In England the ratio of constituencies differed by 40 to 1 before the Redistribution Act 1885 but remained at 7 to 1 until well into the twentieth century [328] .
The numerical size of State electorates at federation
The numerical size of electorates in the States at the time of federation and the terms of s 31 of the Constitution indicate that the view prevailing in England in 1900 also prevailed in Australia and that the makers of our Constitution were content to allow it to continue unless and until the parliamentary representatives declared otherwise. Section 31 of the Constitution provides that, until the Parliament otherwise provided but subject to the Constitution, the laws in force in each State for the time being relating to elections for the more numerous House of Parliament of the State should as nearly as practicable apply to elections in the State of members of the House of Representatives. At the time of federation, the electoral laws of each State did not provide for electorates with equal numbers of voters. On the contrary, those laws permitted very large differences in the number of electors in the electoral divisions in each State. The following were the electoral division ratios applying in elections for various parliaments of the colonies in the period 1897 to 1900: House of Parliament Year of Election Electoral Ratio Legislative Assembly (WA) 1897 38.52 to 1 Legislative Council (WA) 1900 22.22 to 1 Legislative Assembly (NSW) 1898 2.23 to 1 Legislative Assembly (Qd) 1899 5.84 to 1 Legislative Assembly (SA) 1899 8.32 to 1 Legislative Council (SA) 1900 2.11 to 1 Legislative Assembly (Tas) 1900 9.10 to 1 Legislative Assembly (Vict) 1900 3.54 to 1 Legislative Council (Vict) 1900 2.45 to 1
Members of the Legislative Councils of New South Wales and Queensland were appointed to office.
The extent of the franchise in the various colonies at the time of federation also make it highly unlikely that the words "chosen by the people" were intended to imply equality of voting power for individual electors. As I have pointed out, until the federal Parliament legislated, s 30 made the qualifications of electors in the more numerous House of Parliament of a State the qualifications for federal elections for the House of Representatives in that State. There were very large differences in the franchises in each State. In New South Wales, only men were entitled to vote [329] . In Victoria, only men over the age of twenty-one who were natural born British subjects or naturalised subjects of three years standing with at least one of those years spent in the State were entitled to vote. It was also a condition of the entitlement to vote in that State that a voter should own land with an annual value of £10 or hold a lease of land that was rated with an annual value of £25. University graduates and members of certain occupations were also given a right to vote [330] . In Queensland, only male British subjects over the age of twenty-one were entitled to vote. A voter had to be resident in an electoral district or own land in the district valued at £100 or hold a lease or licence of land in the district of a clear annual value of at least £10. While an elector could vote only once in a district, he was entitled to vote in more than one district if he was qualified. Military personnel and police officers were disqualified from voting while aboriginals of Australia, India, China or the South Sea islands had to meet a freehold property requirement [331] . In Western Australia, both men and women were entitled to vote but property qualifications applied [332] . In Tasmania only British subjects over the age of twenty-one were entitled to vote. Property or other qualifications were additional requirements [333] . In South Australia, men and women over the age of twenty-one were entitled to vote but in elections for the Legislative Council only if certain property requirements were met [334] .
These differences in the franchise show that, although the original States were to be represented in the House of Representatives, proportionately to their populations, the Constitution laid down no requirements apart from s 41 as to which members of the population of a State were to elect its representatives. That being so, it seems unlikely that s 24 was concerned with equal representation for equal numbers of voters in each State. Thus s 30 confirms the view that equality of voting power for the adult people of Australia is not a constitutional requirement. The refusal of the Convention delegates to provide for universal suffrage in the Constitution is itself eloquent evidence that such matters as the franchise and the size of electoral divisions were matters for the States until the Parliament legislated.
No implication of equal numbers in electoral divisions
Given the terms of Pt III of Ch 1 of the Constitution and the electoral laws that applied in Australia at the time of federation, it is impossible in my view to conclude that the Constitution intended that either the federal Parliament or, in the absence of federal legislation, a State was required to have equal numbers of electors in the electoral divisions (if any) in each State. In the light of history and context, the obvious purpose of the words "chosen by the people" in ss 7 and 24 was to prevent the use of such devices as the choosing of senators or members by an electoral college, or by nomination as was the case in New South Wales and Queensland with respect to their Legislative Councils [335] .
In so far as any notion of electoral equality is present in the Constitution, it applies only to the Original States in respect of representatives for those States in the House of Representatives and not to individual electorates which may constitutionally consist of an entire State or of electorates whose size and population are determined by the federal Parliament from time to time or, in the absence of legislation, by the parliaments of the respective States. The plain meaning of the Constitution is that it was for the States to determine the size of electoral divisions within the States until the federal Parliament decided to legislate on the question. If and when the federal Parliament decided to legislate, ss 29, 30 and 31 gave it the power to determine the size and nature of the electoral divisions in each State and the nature of the franchise as well as other electoral matters [336] .
Further, the provisions of Pt II of Ch 1 dealing with the Senate, s 121 dealing with new States, s 122 dealing with Territories and above all s 128 dealing with the power to amend the Constitution show that equality of individual voting power is not a fundamental requirement of "representative democracy" under the Constitution. If representative democracy, as understood outside the context of the Constitution, requires equality of electors in electoral divisions, it does not do so under the Constitution. The Constitution makes the federal Parliament the final arbiter on this matter just as it makes the federal Parliament the final arbiter on whether there should be universal suffrage, secret ballot, preferential or proportional voting or first past the post voting [337] .
The plaintiffs' answer to all these textual, contextual and historical considerations is to accept that, for a long period, equal representation for equal numbers was not required by the Constitution but also to assert that Pt III of Ch 1 is subject to the constitutional principle that the people have to be governed by the principles of representative democracy. Consequently, say the plaintiffs, as soon as representative democracy required an equal number of electors in each division, the power of the federal Parliament under s 29 or, in the absence of federal legislation, the power of State parliaments under that section to legislate for the creation of electoral divisions was subject to that requirement of representative democracy.
But if a free-standing principle of representative democracy exists in the Constitution, it only arises from necessity, and an implication arising from necessity "must be limited by the extent of the need" [338] . Since the requirement of an equal number of voters in electoral districts was not seen as necessary for the preservation of the Constitution or its system of government at the time of federation, it is difficult to see how it can be rationally concluded that it is necessary at the present time. If the political principle of representative democracy now requires that one vote have one value, then in so far as that principle is an implied part of the Constitution, its political meaning must give way to the natural and ordinary meaning of ss 24, 29 and 30 of the Constitution and the inferences that arise from ss 7, 121, 122 and 128 of the Constitution.
In the light of ss 29 and 30 of the Constitution, the electoral practices at the time of federation in Great Britain and Australia and those express provisions of the Constitution that require or authorise unequal voting power, it is impossible to hold that the Constitution requires equality of electoral districts.
Representative democracy and equal representation for equal numbers
In any event, I do not think that it is established that in Australia the requirement of equality in the numerical size of electoral districts is so universally conceived as a tenet of representative democracy that the federal Parliament cannot legislate under s 29 of the Constitution without giving effect to it. If it is rationally open to the federal or State Parliaments to conclude that, consistently with representative democracy, other factors concerning parliamentary representation must be weighed against the theory of equal representation for equal numbers, the plaintiffs' case based on representative democracy must fail.
Influenced by United States Supreme Court decisions, many political scientists see representative democracy as requiring equality of voting power [339] . But it is not a view that represents the consensus of opinion in the Australian community. Whether or not representative democracy in Australia requires equality of voting power is a controversial issue in this country [340] . Indeed, it appears that the Australian people do not regard one vote one value as an essential requirement of representative democracy. The issue was put to the people by referendum in 1974 [341] . The "yes" vote attracted 47 per cent of the vote with a majority only in New South Wales [342] . The issue was again put to the people by referendum in 1988 when the people were asked to insert a s [343] into the Constitution. The proposed amendment was overwhelmingly defeated. Every State voted against the referendum. The Australian Capital Territory was the only political unit which supported it. The total "no" vote was 62 per cent [344] .
Further, Canadian courts [345] have rejected the argument that representative democracy in Canada requires equality of voters in electoral divisions for reasons that seem equally applicable to Australia. In Dixon v Attorney-General (British Columbia) [346] , McLachlin J, while Chief Justice of the Supreme Court of British Columbia, said:
The American emphasis on a pure population standard for electoral apportionment may be seen as a product of that country's unique history and conception of democracy. The decisions upholding this standard are heavily influenced by the courts' understanding of the intentions of the framers of the US Constitution
Democracy in Canada is rooted in a different history. Its origins lie not in the debates of the founding fathers, but in the less absolute recesses of the British tradition. Our forefathers did not rebel against the English tradition of democratic government as did the Americans; on the contrary, they embraced it and changed it to suit their own perceptions and needs.
What is that tradition? It was a tradition of evolutionary democracy, of increasing widening of representation through the centuries. But it was also a tradition which, even in its more modern phases, accommodates significant deviation from the ideals of equal representation. Pragmatism, rather than conformity to a philosophical ideal, has been its watchword.
A similar evolutionary tradition has developed in Australia. Although the gap between the largest and smallest electoral divisions in all Australian States has narrowed considerably over the last century, the governments of Australia have consistently legislated for inequalities in representation, inequalities which have generally favoured the non-metropolitan areas [347] . Nevertheless, by 1984, only the electoral divisions of Queensland and Western Australia could be said to be significantly malapportioned [348] . Subsequently, the malapportionment of the Queensland electoral divisions, except for five special districts, has been brought into line with electoral distributions in other States. Even in Western Australia, the gap between the largest and smallest electorates, in terms of the number of electors, has been continually narrowing throughout this century, as the materials before the Court show.
The Australian people and their representatives appear to have accepted Birch's view [349] that:
a political system can properly be described as a system of representative government if it is one in which representatives of the people share, to a significant degree, in the making of political decisions Most students of politics agree that the variety of interests and opinions within a society is too great for them to be adequately represented by any one man. The necessary condition of representative government is therefore said to be the existence of a fair number of representatives of the people, meeting together in some kind of council or assembly. Stephen J expressed a similar view in McKinlay [350] when he said: [R]epresentative democracy is descriptive of a whole spectrum of political institutions, each differing in countless respects yet answering to that generic description. The spectrum has finite limits and in a particular instance there may be absent some quality which is regarded as so essential to representative democracy as to place that instance outside those limits altogether; but at no one point within the range of the spectrum does there exist any single requirement so essential as to be determinative of the existence of representative democracy. 1. Birch, Representative and Responsible Government: An Essay on the British Constitution (1964), pp 13-14. 2. (1975) 135 CLR 1 at 57.
It is no doubt true that something approaching numerical equality of electors within electorates is an important factor, together with much else, in the attainment of what many will regard as representative democracy in its purest form, just as adult suffrage, free of discrimination on the grounds of race, sex, property or educational qualification will likewise aid in its attainment. But neither of these in absolute form is necessarily imported into the Constitution by the selection of representative democracy as the chosen mode of government for the nation. 1. McKinlay (1975) 135 CLR 1 at 57
This statement was made in 1975. The result of the 1988 referendum shows that most Australians still think that representative democracy does not require equal representation for equal numbers notwithstanding that the Constitutional Commission in its Final Report said that it saw "one vote one value [as] an essential principle of democracy" [352] .
No doubt for a time in the nineteenth century, it may have seemed likely that public opinion would ultimately accept that equal representation for equal numbers was necessary if representative government was to be fully achieved. But the fact that that theory did not become the consensus in England and Australia is probably due to the growth of the party system. Once the party system became firmly established, voters' loyalties were to the parties and their leaders rather than to individual members of Parliament [353] . Not equality of voting power but the extent to which a political party's vote translates into seats in Parliament is now seen by many political scientists as the surest guide to the fairness of a particular political system [354] . Equality of numbers in electorates does not guarantee representative government. Because of the party system, electorates can be gerrymandered notwithstanding that they meet the equality standard [355] . Of course, that does not mean that equality of numbers in electoral divisions is not extremely important if an electorate is divided into divisions.
Conversely, inequality in the numerical sizes of electoral divisions does not necessarily mean a party is unfairly represented. Even when electoral divisions are skewed in the manner in which the electoral divisions are skewed in Western Australia, the number of seats obtained by the major political parties frequently represent a rough approximation of the first preference vote of that party. Thus in the 1993 Western Australian election in the Legislative Assembly the Liberal Party gained twenty-six seats, the Labor Party twenty-four seats, the National Party six seats, the Greens (WA) and Australian Democrats no seats and an Independent one seat. The Liberal Party gained 44.15 per cent of the first preference votes and 45.61 per cent of the seats; the Labor Party gained 37.08 per cent of the vote and 42.11 per cent of the seats. Thus, although the electoral distribution in Western Australia is weighted against the metropolitan seats where the Labor Party might be expected to have the largest share of its vote, it gained a greater percentage of seats than its share of the vote. This was achieved even though the National Party which gained 5.31 per cent of the vote obtained 10.53 per cent of the seats. However, the Greens (WA) which obtained 4.31 per cent of the vote obtained no seats [356] . If a Statewide proportional representation system had been in use, the Liberal Party would have gained 47.4 per cent of the seats, giving that party an additional seat. The Labor Party would have gained 40.3 per cent of the vote, giving them one less seat. The National Party would have received 5.3 per cent of the seats, giving them three less seats. The Greens (WA) would have obtained 5.3 per cent, giving them three seats. The Australian Democrats would have gained 1.7 per cent of the seats giving them one seat [357] . A fairly similar result would have obtained by applying a Statewide proportional representation system to the Legislative Council election. This would have resulted in the Liberal Party gaining one seat, the Labor Party losing one seat, the National Party losing two seats and the Greens (WA) and the Australian Democrats gaining one seat each. Under the present system operating in Western Australia, the beneficiaries would seem to be the Labor Party and the National Party, notwithstanding that the Labor Party's voters are probably concentrated in the Metropolitan Area while the National Party's voters are concentrated in the non-metropolitan area. The losers are the Liberal Party, the Greens (WA) and the Australian Democrats. A Statewide proportional system rather than equality of numbers in electoral divisions would seem necessary to obtain equality of representation in the Parliament of Western Australia because, whatever the size of those divisions, the minor parties seem unlikely to obtain their fair share of parliamentary representation.
Thus, the way that a party's voters are dispersed through the electorate is likely to have a greater effect on representation in the Houses of Parliament than is equality of electoral districts. In the 1990 federal election, for example, the Labor Party won 52.7 per cent of the seats in the House of Representatives (seventy-eight out of 148), although it only obtained 49 per cent of the national two party preferred vote [358] .
Only if the Court concluded that equal representation for equal numbers was an indispensable element of representative democracy would it be justified in holding that the parliaments of Australia cannot legislate inconsistently with that element. Notwithstanding the views of eminent political scientists and the Constitutional Commission, a very substantial number of Australian electors, perhaps a majority, do not see equal representation for equal numbers as indispensable to representative government or democracy in Australia. When such a division of opinion occurs about what is essentially a political question, this Court should be slow to substitute its views for that of the Parliaments as to what representative democracy requires.
However, for the reasons I have given, the Constitution does not require equal representation for equal numbers in elections in a State for the House of Representatives whatever the principles of representative democracy may currently require.
If the Constitution does not require the electoral divisions for federal elections to be equal in numerical size, there is no ground for supposing that the Constitution requires that the electoral divisions for State elections be equal. In Australian Capital Television, Theophanous and Stephens , the majority justices found that s 128 of the Constitution were the basis for an implication that the principle of representative democracy is part of the Constitution. If the principle of representative democracy is based on Ch 1 and s 128 of the Constitution and applies to State Constitutions, it can surely only do so through the operation of s 106 of the Constitution [359] . However, in Theophanous and Stephens the majority justices seem to have seen the principle as an overarching and free-standing provision equivalent to a hypothetical s 129 of the Constitution. But even if this is the correct view of the implied constitutional principle of representative democracy after those two cases, since it does not require equal representation for equal numbers in federal elections, it surely cannot shake off its foundations and require it in State elections.
No doubt the logic of the reasoning in Theophanous and Stephens would seem to imply that the principle of representative democracy applies generally throughout the Constitution and could require equality of electorate divisions for State elections even though other provisions of the Constitution demonstrate that such equality is not required in federal elections. In other words, the express provisions of the Constitution may confine representative democracy in its application to federal elections but, in the absence of some provision in the Constitution, the principle is not so confined in its application to State elections. If the logic of Theophanous and Stephens requires this result, it provides the strongest ground for overruling those decisions as soon as possible. If that is the logic of those decisions, their continued operation can only distort the meaning and application of the Constitution.
In my opinion, once it is held that the system of representative government set up by the Constitution does not require equal representation for equal divisions in federal elections, it must follow that the Constitution imposes no such obligation in State elections. It follows that the plaintiffs' claim based on the Constitution must be rejected.
The plaintiffs assert, however, that independently of the Constitution, the Western Australian Constitution contains a principle of representative democracy which requires that elections to the Legislative Assembly and the Legislative Council be based on electoral districts and regions that contain an equal number of electors. They contend that this requirement of electoral equality is "more directly recognised and reflected in s 73(2)(c) of the [Western Australian Constitution]" which is a "manner and form" provision within the meaning of s 6 of the Australia Act 1986. Consequently, so the plaintiffs contend, the 1987 Amending Act is invalid because it was a law to which s 73(2)(c) applied and had not been submitted to the electors in accordance with that paragraph.
The effect of s 73 of the Western Australian Constitution was summarised by Mason CJ and Toohey and Gaudron JJ in Stephens [360] as follows:
Section 73 of the Constitution Act 1889 WA provides that the legislature of the State "shall have full power and authority, from time to time, by any Act, to repeal or alter any of the provisions of this Act". However, it is not lawful to present to the Governor for assent by the Queen [361] : "any Bill by which any change in the Constitution of the Legislative Council or of the Legislative Assembly shall be effected, unless the second and third readings of such Bill shall have been passed with the concurrence of an absolute majority of the whole number of the members for the time being of the Legislative Council and the Legislative Assembly respectively."
Section 73(2) was inserted into the Constitution by an amendment enacted in 1978. That sub-section further restricts the capacity of the legislature to enact a Bill which expressly or impliedly provides for the abolition of either the Council or the Assembly (s 73(2)(b)) or provides that either House "shall be composed of members other than members chosen directly by the people" (s 73(2)(c)). The sub-section requires that such a Bill should not be presented for assent unless "the second and third readings of the Bill shall have been passed with the concurrence of an absolute majority of the whole number of the members for the time being of the Legislative Council and the Legislative Assembly, respectively" (s 73(2)(f)) and the Bill has been approved by a majority of electors of the State at a referendum. [362]
It is not open to doubt, in my opinion, that s 73(2)(c) is a "manner and form" provision within the meaning of s 6 of the Australia Act 1986. Nor is it open to doubt that the Amending Act of 1987 was one "respecting the constitution, powers or procedure of the Parliament of the State" within the meaning of s 6 of the Australia Act or that the Amending Act of 1987, not having been submitted to a referendum as required by s 73(2), is invalid if it "expressly or impliedly provide[d] that the Legislative Council or the Legislative Assembly shall be composed of members other than members chosen directly by the people". However, it does not follow that, because the Court in Stephens saw s 73(2) as reinforcing representative democracy in Western Australia, s 73(2)(c) requires equality of numbers in the electoral districts and regions of Western Australia.
The plaintiffs contend that the words "chosen directly by the people" necessarily import into the Western Australian Constitution a requirement of equality of electoral districts and regions. However, as I pointed out when discussing the similar words of s 24 of the Constitution, the ordinary and natural meaning of those words says nothing about how the members of the Legislative Council or Legislative Assembly are to be elected. If those words have the meaning for which the plaintiffs contend, it can only be because they were understood in that sense in Western Australia when s 73(2) was enacted. But not only is there no indication that the Western Australian Parliament understood the words in that sense, there are clear indications to the opposite effect.
When s 73(2)(c) was enacted, Western Australia was divided into electoral provinces and districts not dissimilar from those that resulted from the 1987 Amending Act. It is unnecessary to set out the details. It is sufficient to say that the relevant legislation authorised and resulted in considerable disparities in the number of electors who voted in the various districts and provinces of Western Australia. In the history of Western Australian electoral distributions there was nothing unusual about these disparities. Very significant disparities in the numbers of voters in electorates have always been a marked feature of electoral distributions in Western Australia. In the last election before federation, the electoral disparities were as high as 38.52 to 1. Given this history, it would be surprising indeed if the words "the Legislative Council [and] Legislative Assembly shall be composed of members chosen directly by the people" (s 73(2)(c)) were directed to equal representation for equal districts and provinces.
Thus, given the disparities in the districts and provinces at the time that s 73(2)(c) was enacted, the most likely conclusion is that the Parliament of Western Australia intended that the words "composed by members chosen directly by the people" meant what they said. They required the members to be elected by the direct vote of the people as opposed to being appointed by the Crown or the Ministry or being elected by electoral colleges. That that was the intention of the Parliament was made plain by the Premier in the course of the parliamentary discussion when he said that one of the objectives of s 73(2)(c) was to prevent any attempt "to by-pass the right of the electors at large to elect the members of either House" [364] . That s 73(2)(c) was not concerned with the size of electorates or electoral quality was made plain in the Legislative Council. The Second Reading Speech expressly stated that s 73(2)(c) did not affect electoral boundaries [365] .
In addition, three years earlier this Court had construed the similar words "directly chosen by the people" and had held that they did not import any notion of equal representation for equal groups of electors or of one vote one value [366] . It is a reasonable assumption that the words of s 73(2)(c) were chosen having regard to the interpretation placed on the similar words in s 24 of the Constitution in McKinlay .
In my opinion, having regard to the history of electoral divisions and provinces in Western Australia, the nature of those divisions and provinces in 1978, the decision in McKinlay and the statements in the Legislative Council and the Legislative Assembly, it is impossible to hold that s 73(2)(c) requires that a bill affecting electoral districts and regions must be submitted to a referendum to be valid unless it provides for the numerical equality of those districts or regions.
Once that conclusion is reached, the existence of any implied principle of representative democracy in the Western Australian Constitution is irrelevant. If s 73(2)(c) does not require equality of numbers in electoral districts or regions in Western Australia, then the Western Australian Constitution has no entrenched provision that prevents the Parliament of Western Australia from legislating for unequal numbers in electoral districts. It is legally and constitutionally irrelevant that such legislation is inconsistent with the principles of representative democracy. The Parliament of Western Australia has the constitutional power to alter the principles of representative democracy or legislate inconsistently with those principles unless it is prevented from doing so by an entrenched provision in the Western Australian Constitution or some provision in the Commonwealth Constitution. Because neither Constitution contains a provision that prevents the Parliament of Western Australia from legislating for unequal numbers in electoral divisions and regions, that Parliament may modify or legislate inconsistently with the principles of representative democracy in so far as those principles require equality of numbers.
The claim of the plaintiffs based on the Western Australian Constitution must be rejected.
Questions 1 and 2 of the case stated should be answered, No. It is unnecessary to answer Question 3.
The first and second plaintiffs, Mr McGinty and Mr Gallop, are members elected to the Legislative Assembly of the Parliament of the State of Western Australia and the third plaintiff, Mr Halden, is a member of the Legislative Council of that Parliament. By their action commenced in the original jurisdiction of this Court, they seek declaratory relief as to the invalidity of certain legislation of Western Australia. They complain that, in respect of voting for representatives to either chamber of the Parliament, the legislation brings about a gross malapportionment in favour of non-metropolitan voters.
The matter comes before this Court upon a case stated pursuant to s 18 of the Judiciary Act 1903 Cth.
A number of recent decisions in this Court, Nationwide News Pty Ltd v Wills [367] , Australian Capital Television Pty Ltd v The Commonwealth [368] , Theophanous v Herald & Weekly Times Ltd [369] , Stephens v West Australian Newspapers Ltd [370] , and Cunliffe v The Commonwealth [371] , turned upon or involved some consideration of the consequences, for validity of federal and State statutes and for the continued operation of the common law, of recognition that the federal Constitution and that of Western Australia establish and maintain systems of representative government. The subject matter of the statute law and common law at issue in those decisions was regulation or prohibition with criminal or civil sanction of the publication or dissemination of certain words and visual images. At stake was held to be the efficacious working of the system of representative government. What was not required for these decisions, but is necessary now, is a fuller consideration of those characteristics or elements of an electoral system which must be present to answer a constitutionally mandated system of representative government. In Stephens it was assumed that in Western Australia presently there was in operation such a system [372] . This assumption the plaintiffs now dispute.
The effect of amendments made to the Electoral Distribution Act 1947 WA (the 1947 Act) by the Acts Amendment (Electoral Reform) Act 1987 WA (the 1987 Act) [373] upon representation in the Legislative Assembly has been the creation of electoral districts within the metropolitan zone based on Perth of approximately 23,000 electors plus or minus 15 per cent per electoral district, and in the non-metropolitan zone of approximately 12,000 electors plus or minus 15 per cent. Thus, whilst there is relative equality within the two zones, there is marked difference between them.
Changes made by ss 6 and 7 of the 1987 Act to the Constitution Acts Amendment Act 1899 WA (the 1899 Act) altered the structure of the Legislative Council by providing for six multi-member "electoral regions", three in the metropolitan zone and three in the non-metropolitan zone. The plaintiffs complain that, despite the concentration of the majority of the electors in the metropolitan zone, the distribution of seats is spread evenly between the two zones.
Section 83 of the 1987 Act inserted a new Schedule (headed "Counting Of Votes At Legislative Council Elections") in the Electoral Act 1907 WA. A system of counting the votes known as the "single transferable vote" is used. A successful candidate for election to the Legislative Council must receive a particular quota of the total valid votes cast. This may be illustrated as follows. In the 1993 general election the quota for the East Metropolitan Region was 32,822. This was determined by dividing the total number of valid votes cast in that region (196,927) by the number of members to be elected (five) plus one, and then adding one vote [374] .
At the last general election, held on 6 February 1993, the ratio of the highest to the lowest quota for election of a member of the Legislative Council, on the basis of the regions, was 3.76:1. The ratio between the largest number of electors in a single region of the Legislative Council and the smallest number of electors in another such region, when divided by the number of members in each of those regions, was 3.42:1. The ratio between the largest number and the smallest number of voters enrolled in each district for the Legislative Assembly at the 1993 general election was 2.91:1.
Put in short form, the plaintiffs' submissions are as follows. First, they contend that, even though a requirement of "one vote one value" is not found in the text of the Constitution, an essential characteristic of the system of representative government established by the Constitution is a franchise conferring "one vote one value". This is to be achieved by the presence of "practical equality", or something close to it, in numbers of electors in electoral divisions. Then it is said that a law prescribing and regulating the federal franchise must not depart from that requirement. Further, the plaintiffs submit that, even if otherwise not an essential characteristic as described, this principle of "one vote one value" is implicit in the very term "directly chosen by the people" in ss 7 and 24 of the Constitution. As the final step, they contend that, by force of the Constitution, the laws of Western Australia which prescribe and regulate the franchise at the State level for each chamber of the legislature must satisfy (but fail to do so) the same requirement of "one vote one value" as is mandated by the Constitution for the laws of the Commonwealth.
In 1975, in Attorney-General (Cth); Ex rel McKinlay v The Commonwealth [375] , it was held (Barwick CJ, McTiernan, Gibbs, Stephen, Mason and Jacobs JJ; Murphy J dissenting) that s 24 of the Constitution does not require the number of people or the number of electors in electoral divisions for the House of Representatives to be equal. To the extent that this authority bars their path, the plaintiffs seek that it be overruled. They are, somewhat unusually, joined in that endeavour to overthrow a decision in favour of the Commonwealth by the Attorney-General for the Commonwealth, who is an intervener. The Attorney's submission appears to be that the present legislation of the Commonwealth (ss 59 and 73 of the Commonwealth Electoral Act 1918 Cth (the 1918 Act)) sufficiently observes a requirement of "one vote one value", in the sense contended for by the plaintiffs and that this is not merely the work of a benign legislature, but observance of a constitutional imperative.
Alternatively, the plaintiffs rely upon the alleged failure of the 1987 Act to comply with what is said to be the requirement for prior approval by the electors of the State. That requirement was imposed by amendment in 1978 [376] to s 73 of the Constitution Act 1889 WA (the 1889 Act). The requirement applies (s 73(2)(c)) to a Bill which "expressly or impliedly" provides that either House shall be composed of members other than members "chosen directly by the people". The insertion of sub-s (2) into s 73 was made after the construction given to s 24 of the federal Constitution in McKinlay . Hence, the importance to the plaintiffs, on each branch of the case, in overcoming McKinlay . The plaintiffs submit that the 1987 Act is invalid by reason of contravention of s 73(2) of the 1889 Act and they further submit that the 1987 Act amounted to an attempt to alter the Constitution of the State other than in accordance with that Constitution. In that latter regard, they rely upon s 106 of the federal Constitution.
It may be said that, upon a consideration of any one constituency established under the present Western Australian electoral structure, no one elector has a vote of greater "value" than that of any other elector in that constituency. However, the complaint of the plaintiffs stems from the disparity in the number of electors between constituencies.
Their submissions reflect what was said by Powell J in Davis v Bandemer [377] :
The concept of "representation" necessarily applies to groups: groups of voters elect representatives, individual voters do not. Gross population disparities violate the mandate of equal representation by denying voters residing in heavily populated districts, as a group, the opportunity to elect the number of representatives to which their voting strength otherwise would entitle them. While population disparities do dilute the weight of individual votes, their discriminatory effect is felt only when those individual votes are combined. Thus, the fact that individual voters in heavily populated districts are free to cast their ballot has no bearing on a claim of malapportionment. 1. (1986) 478 US 109 at 167.
In the Legislative Assembly, a consequence of this disparity may be that one representative has required for election the support of more voters than that required for the successful representative in another constituency. Governments are formed from one political party or a coalition of parties so as to command the support of the lower house. The result of this disparity may be (although the State has provided statistics suggesting that this by no means always is so) that to a significant degree the majority comprising members returned who support one major party or a coalition of parties does not correspond to the overall party vote across the State. That state of affairs also may reflect a range of other circumstances, for example, the operation of preferential voting and the failure of each party to field candidates for each constituency. These matters illustrate the point brought out in argument that the significance of the notion of "one vote one value" reflects the concerns of the established political parties as much as any doctrinal solicitude for the true and proper nature of the franchise.
No mention was made in the text of the Constitution of political parties until the substitution of the new s 15 in 1977. Unstable factions had been a significant feature of colonial parliamentary government. For example, in New South Wales, in the years between 1856 and 1887, twenty-three ministries had been formed with an average life of a little over 16 months [378] . At the time of the adoption of the Constitution these factions were yielding to modern party political structures, with national organisations. Performance of the task of fashioning fully a system of representative self-government, left by many provisions of the Constitution to future Parliaments, fell to legislators subjected to the pressures of party politics. The interests of that party political system and of the established parties in a sense have come to be identified with those of representative government itself.
Experience in the United States demonstrates that to insist, on constitutional grounds, upon equal numbers of electors in constituencies does not necessarily avoid any skewing of the overall party vote away from party representation in the legislature. In Davis v Bandemer [379] , the majority of the Supreme Court of the United States [380] held that a claim that the boundaries of the constituencies for a bicameral State legislature were distorted so as seriously to dilute or eliminate the voting power of electors affiliated with a particular party raised an issue of equal protection under the Fourteenth Amendment. The Supreme Court rejected the submission that this was a political question and non-justiciable. Other decisions to like effect deal with issues of "racial gerrymander" [381] . Recently, a more restrictive view has been taken as to what is requisite for standing to institute such cases [382] .
In the present case, the plaintiffs eschew any claim of "gerrymander". They rest upon that particular meaning of the principle of "one vote one value" which I have sought to describe above.
Before giving further consideration to the particular legislative provisions whose validity the plaintiffs attack, it is appropriate to refer briefly to the legislative components of the present Western Australian Constitution.
The Imperial, colonial and State legislation together comprising the written provisions of the Western Australian Constitution, as it then stood, were collected and analysed by Wilson J in Western Australia v Wilsmore [383] . Then, as now, it was unnecessary to decide whether the Constitution of the State, particularly for the purposes of s 106 of the Australian Constitution, also includes other prerogative instruments or statutory and common law rules and principles which define and regulate the legislative, executive and judicial elements of government [384] . Wilson J described the 1889 Act (which was the scheduled Bill to the Imperial statute, the Western Australian Constitution Act 1890 Imp) [385] as "[t]he keystone of the present Constitution of Western Australia " [386] . Further provision was made by the Constitution Act Amendment Act 1893 WA but this in turn was repealed by the 1899 Act. The present case principally concerns the construction of the additions to s 73 of the 1889 Act made by s 6 of the 1978 Act. Wilsmore dealt with s 73 in its earlier form [387] .
The constitutional changes made in Western Australia in the last decade of the nineteenth century took place at a time of rapid growth in population, from 48,500 in 1890 to 180,000 by 1900 [388] . At that time neither enrolment nor voting was compulsory. The method of counting votes was that known as "first past the post". In 1893, property qualifications for candidates and voters for the Legislative Assembly were abolished. Women received franchise in 1899 but until 1920 remained ineligible to stand as candidates and, as with men, a property qualification was required for exercise of the franchise in elections for the Legislative Council. In elections for the Council, property owners received one vote for each region in which property of the requisite value was situated. The property qualification and plural voting ordained by s 15 of the 1899 Act were finally abolished by s 8 of the Constitution Acts Amendment Act (No 2) 1963 WA. The right to vote at State elections was conferred on persons on attaining the age of eighteen years by the Electoral Act Amendment Act (No 2) 1970 WA.
From the commencement of representative government in Western Australia, there were considerable disparities between the number of electors in various constituencies. At the election held for the Legislative Assembly in 1897, of the forty-four constituencies, each of which returned one member, one had fifty-four electors, twenty-seven had less than 500, six had over 1,000 and only one had over 2,000. At the election held in 1900 for the Legislative Council, of the eight constituencies into which the colony then was divided, each of which was represented by three members serving staggered six-year terms, one had just 156 electors, whilst the Metropolitan-Suburban constituency had 3,467 electors.
One of the themes in the submissions to this Court for the State of Western Australia was that the expansion over the last century in the economy and population of Western Australia has been accompanied by legislative changes in the operation of the franchise for both chambers of the Parliament. The consistent tendency of these changes has been to lessen the imbalance between the metropolitan and non-metropolitan vote. Nevertheless, so it is submitted by the State, even though one may accept the concept of "equality of voting power" as understood by the plaintiffs, that cannot be the sole factor to be taken into account in the provision of representative government. Factors such as the geographical expanse of the State, variations in its topography, community interests, difficulty in ready communication between legislators and constituents and the like are said to be significant, and to be given appropriate significance in the legislation under attack, particularly by s 7 of the 1947 Act.
To further consideration of this legislation I now turn.
Section 6 of the 1987 Act inserted a new s 5 in the 1899 Act. Section 5 of the 1899 Act now states:
The Legislative Council shall consist of 34 elected members who shall be returned and sit for electoral regions as defined under section 6. Section 7 of the 1987 Act also inserted a new s 6 in the 1899 Act. Section 6 of the 1899 Act now provides:
(1) The State shall be divided into 6 electoral regions under the Electoral Distribution Act 1947.
(2) The electoral regions known, respectively, as the North Metropolitan Region and the South West Region shall each return 7 members to serve in the Legislative Council.
(3) The electoral regions known, respectively, as the South Metropolitan Region, the East Metropolitan Region, the Agricultural Region and the Mining and Pastoral Region shall each return 5 members to serve in the Legislative Council.
Sections 18 and 19 of the 1899 Act now provide that the Legislative Assembly shall consist of fifty-seven members, each returned by an electoral district provided under the 1947 Act.
The 1987 Act made significant changes to the relevant provisions of the 1947 Act. Section 89 repealed s 2A of the 1947 Act and inserted a new s 2A. Section 2A(2), in the events that since have happened, requires the State to be divided into districts and regions in accordance with the 1947 Act. Section 92 substituted a new s 6. This requires the Electoral Distribution Commissioners to divide the Metropolitan Area (an expression defined in a new s 1A inserted by s 87 of the 1987 Act) into thirty-four districts (s 6(1)(a)) and to divide the area comprising the remainder of the State into twenty-three districts (s 6(1)(b)). The new s 6(2) should be set out in full. It states:
The Commissioners shall make the division of an area mentioned in subsection (1)(a) or (b) into districts in accordance with the principle that the number of enrolled electors comprised in any district in the area must not be more than 15% greater, or more than 15% less, than the quotient obtained by dividing the total number of enrolled electors in the area by the number of districts into which the area is to be divided. Section 7 of the 1947 Act (also inserted by s 92 of the 1987 Act) specifies various matters to be given "due consideration" by the Commissioners in making the division of the State into regions and districts. These include "community of interest" [389] , "means of communication and distance from the capital", "physical features", "existing boundaries of regions and districts", "existing local government boundaries" and "the trend of demographic changes". 1. Section 92 of the 1987 Act inserted the words "community of interests". However, the words in the 1947 Act, reprinted as at 26 November 1987, are "community of interest". Under s 8 of the Reprints Act 1984 WA, pursuant to which the 1947 Act was reprinted, the reprinted Act should be taken to be correct.
Section 94 of the 1987 Act inserted a new s 9 into the 1947 Act. This now states:
The Commissioners shall divide the State into 6 regions so that -
(a) 3 regions, to be known, respectively, as the North Metropolitan Region, the South Metropolitan Region and the East Metropolitan Region, each consist of complete and contiguous districts that together form the Metropolitan Area;
(b) one region, to be known as the Mining and Pastoral Region, consists of complete and contiguous districts that are remote from the capital and where the land use is primarily for mining and pastoral purposes;
(c) one region, known as the Agricultural Region, consists of complete and contiguous districts that together form an area that is generally south, or south and west, of and adjacent to the Mining and Pastoral Region; and
(d) the remaining region, to be known as the South West Region, consists of complete and contiguous districts.
The plaintiffs claim a declaration that s 6 of the 1899 Act and ss 2A(2), 6 and 9 of the 1947 Act are invalid. More precisely, as I would see it, their allegation is that the provisions of the 1987 Act which repealed the previous provisions of the 1947 Act and the 1899 Act are invalid. That is to say, the plaintiffs' complaint is directed against ss 7, 89, 92 (in so far as it repealed the previous s 6 of the 1947 Act and inserted the new s 6) and 94 of the 1987 Act.
However, the questions which have been reserved on the case stated for the Full Court are:
(i) Is s 6 of the Constitution Acts Amendment Act 1899 WA invalid?
(ii) Are ss 2A(2), 6 and 9 of the Electoral Distribution Act 1947 WA, or any of them, invalid?
(iii) If any of ss 2A(2), 6 or 9 of the Electoral Distribution Act 1947 WA is or are invalid, is it or are they severable and, if so, to what extent?
The plaintiffs place considerable reliance upon what they submitted was authoritative guidance to be found in a series of decisions of the Supreme Court of the United States delivered over the past 30 years. The series commenced in 1962 with Baker v Carr [390] in which the Supreme Court held that the challenge by the plaintiffs to the Tennessee apportionment legislation of 1901 was within federal jurisdiction and presented a justiciable claim. Previous authority had treated the right to vote, even at federal level, as derived from the States [391] .
The plaintiffs contend that Baker v Carr and the succeeding decisions give effect to a constitutional principle of equality of voting power. But, upon examination, they provide but limited assistance to the plaintiffs in the present litigation.
At the federal level in the United States, the focus of attention has been congressional districts and the statement in Art 1, s 2 of the Constitution:
The House of Representatives shall be composed of Members chosen every second Year by the People of the several States, and the Electors in each State shall have the Qualifications requisite for Electors of the most numerous Branch of the State Legislature. (Emphasis added.) Section 4 of Art 1 provides that: [t]he Times, Places and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof; but the Congress may at any time by Law make or alter such Regulations, except as to the Places of chusing Senators. [392] 1. See US Term Limits Inc v Thornton (1995) 131 LawEd 2d 881, which holds that the States may not impose limits on the length of time a representative may serve in Congress.
At the State level, the central provision has been the Equal Protection Clause in s 1 of the Fourteenth Amendment made to the Constitution in 1868. This states:
All persons born or naturalized in the United States, and subject to the jurisdiction thereof, are citizens of the United States and of the State wherein they reside. No State shall make or enforce any law which shall abridge the privileges or immunities of citizens of the United States; nor shall any State deprive any person of life, liberty, or property, without due process of law; nor deny to any person within its jurisdiction the equal protection of the laws. (Emphasis added.)
Before turning to the decisions, two points should be made. First, the concept or doctrine of representative democracy operates in the United States without compulsory registration of electors or compulsory voting by those so registered, yet at State and local level the range and variety of public offices filled by periodic popular election is far more extensive than has ever been the case in Australia. In the United States, this expands the need for restraint upon legislative interference with freedom of communication on matters relevant to political discussion.
Secondly, the difference between the constitutional base provided at federal and State level, on the one hand, by Art 1, s 2 and, on the other, in the Fourteenth Amendment has contributed to differential treatment by the Supreme Court of the requirement of equality of voting power.
In Wesberry v Sanders [393] , the Supreme Court pointed out that, as late as 1842, seven States still conducted congressional elections with the State voting as one electorate, so that in such States each vote was worth as much as any of the others; turning to the modern system of division into districts, the Court required the States to draw their congressional districts so that "as nearly as is practicable one man's vote in a congressional election is to be worth as much as another's". Then, in Kirkpatrick v Preisler [394] , the Supreme Court held that, in drawing congressional districts for the House of Representatives, a State must "make a good-faith effort to achieve precise mathematical equality". What this meant then became apparent in White v Weiser [395] . There the Supreme Court rejected a plan in Texas for 24 districts in which the average deviation of all districts from the ideal was 0.745 per cent, the largest district exceeded the ideal by 2.43 per cent and the smallest district was under the ideal by 1.7 per cent [396] .
In the present case, neither the plaintiffs nor those interveners who supported them, including the Commonwealth, advocated any requirement of mathematical precision in the sense demonstrated in the requirements for congressional districts. Indeed, the provision in federal law, s 73 of the 1918 Act, permitting a margin of allowance whereby the quota may be departed from up to one-tenth more or one-tenth less would be unacceptable in the United States.
In the United States at the State level, the principle of one person one vote, flowing from the equal protection clause, does not prevent the election of the governor of a State by the State legislature rather than by direct popular election [397] . Nor is the election of State or local judges subjected to the principle of one person one vote; this is on the footing that, unlike the legislative function of government, the judicial function is to serve the people not by representing them but by administering the law for the benefit of all [398] .
Otherwise, whenever a State or local government decides to select persons by popular election to perform "governmental functions", the Fourteenth Amendment requires giving to each qualified voter an "equal opportunity" to participate in that election; when members of an elected body are chosen from separate districts, each must be established on a basis that will ensure, so far as is practicable, that equal numbers of voters can vote for proportionally equal numbers of officials [399] . However, it appears that, whilst population alone is the primary criterion, a broader latitude is afforded to the States under the equal protection clause than that which applies at federal level. For example, multi-member legislative districts may be permitted if there are "persuasive justifications" [400] . The result of the decisions upon the 14 Amendment is summarised as follows [401] :
When the Supreme Court examines population differences in voter districts, the Court will compare the size of the most and least populous districts with the theoretically "ideal" district. For example, if a city has 5 million people and a city council composed of 5 city council persons, each of whom is elected from a separate geographic area of the city, the ideal population of each district would be 1 million persons. Let us assume that in our hypothetical city the 5 voting districts have populations of: 600,000; 800,000; 1,000,000; 1,200,000; 1,400,000. It would be common for courts (though not mathematicians) to describe the most and least populous districts as each having a 40% variance from the ideal district and the city as a whole having a total variation of 80%. [T]he Court has found that a state or local legislative map that has a total deviation of 10% or less does not require special justification by the government. The Court has upheld state and local legislative maps with total deviations of almost 20%, and it allowed one state to provide a guarantee of at least one representative in the legislature for each county. 1. Hadley v Junior College District (1970) 397 US 50The plaintiffs were residents and taxpayers of the Kansas City School District and elected officials were trustees who conducted and managed the affairs of that District. 2. Chapman v Meier (1975) 420 US 1 at 26-27See also Davis v Mann (1964) 377 US 678 at 686 Mahan v Howell (1973) 410 US 315 at 325 Connor v Finch (1977) 431 US 407 at 415 3. Rotunda and Nowak, Treatise on Constitutional Law 2nd ed, (1992), vol 3, s18.36, p 452.
In Mahan v Howell [402] , the Supreme Court upheld a legislative apportionment plan of Virginia with a total percentage deviation of 16.4 per cent, one district being over represented by 6.8 per cent and another under represented by 9.6 per cent. On the other hand, in Connor v Finch [403] , the Supreme Court held that a Mississippi legislative map with maximum deviations of 16.5 per cent for the districts for the upper house and 19.3 per cent for the lower house districts were invalid in part because there had been no demonstration of unique features of the political or governmental structure of the State which would have justified this deviation from voter equality. Finally, in Brown v Thomson [404] , the Supreme Court upheld a system in Wyoming which guaranteed a representative in its legislature to all twenty-three counties, including the least populous county in the State, which had a population of 2,924, whilst the ideal county had a population of 7,337.
The effect of these decisions has been described in Rotunda and Nowak, Treatise on Constitutional Law [405] in these terms:
There will be fewer legitimate reasons for population variances between congressional districts than for variations in state or local legislative districts. States may have more legitimate reasons for wishing to keep voter groups in county or other political subdivisions when voting for state legislative positions. Such political divisions often have political powers as units of local government. States may also wish to guarantee representation to small counties because the state legislative system may act on matters which clearly affect different counties in different ways and in which legislative input from all counties or political subdivisions is important. 1. 2nd ed, (1992), vol 3, s18.36, p 449.
If regard is had in this way to the effect of the principle of equality of voting power at all levels in the United States, it will be seen that the result does not favour the present plaintiffs to anything like the degree their submissions suggested.
Further, there are more fundamental reasons why no great weight should be placed upon developments in the United States case law over the last thirty years. Those reasons were, to my mind and with respect, cogently expressed by McLachlin J in her judgments in Dixon v Attorney-General (British Columbia) [406] , when her Ladyship was Chief Justice of the Supreme Court of British Columbia, and as a member of the Supreme Court of Canada in Reference re Electoral Boundaries Commission Act [407] .
In the first of these cases [408] , her Ladyship summed up the position as follows:
In short, the development of Canadian democracy has different roots and has followed a different course than has the American concept of democracy. The rationalist ideals embodied in the French revolutionary slogan "Liberty, equality and fraternity" dominated the thinking of the founding fathers of the American nation. Whatever the actual practice (slaves did not win the right to vote until 1870), the ideal of equal representation was present from the beginning in the United States. The fact that democracy in the United States sprang from a revolution ushering in a new regime, meant that at least in theory the ideals of equality can be seen as having been embraced from the outset in an absolute fashion, something that never occurred either in England or Canada, where gradual change and accommodation of other factors has been the norm. The historical and philosophical tradition upon which the post-1962 "One Man, One Vote" decisions in the United States are grounded, is absent in Canada. The American jurisprudence cannot be transported into the Canadian legal context without careful consideration and, where necessary, modification. 1. Dixon (1989) 59 DLR(4th) 247 at 263.
Earlier in the judgment, after discussion of various of the United States decisions, it was said [409] :
The American emphasis on a pure population standard for electoral apportionment may be seen as a product of that country's unique history and conception of democracy. The decisions upholding this standard are heavily influenced by the courts' understanding of the intentions of the framers of the US Constitution; indeed in Wesberry v Sanders [410] , the opinion of the majority of the Supreme Court focused almost exclusively on the debates and values related to pre-constitutional conventions and discussions. Moreover, the recent insistence on voting equality in its pure form may be seen in part, at least, as a reaction to marked and entrenched voting disparities, reflecting an acute departure from the ideals espoused by the founding fathers. In Baker v Carr [411] , the disparities were as high as 19 to 1 and in Reynolds v Sims [412] up to 41 to 1, being based on the 1900 census despite massive intervening population growth and urbanisation.
Democracy in Canada is rooted in a different history. Its origins lie not in the debates of the founding fathers, but in the less absolute recesses of the British tradition. Our forefathers did not rebel against the English tradition of democratic government as did the Americans; on the contrary, they embraced it and changed it to suit their own perceptions and needs.
What is that tradition? It was a tradition of evolutionary democracy, of increasing widening of representation through the centuries. But it was also a tradition which, even in its more modern phases, accommodates significant deviation from the ideals of equal representation. Pragmatism, rather than conformity to a philosophical ideal, has been its watchword.
In Dixon [413] , reference is made to the position as it then stood in Australia after McKinlay . The Canadian decisions were concerned with the application of the terms of s 3 of the Canadian Charter of Rights and Freedoms, "[e]very citizen has the right to vote in an election of members of the House of Commons or of a legislative assembly". The result has been to reject the contention that this requires absolute, or as near as practicable to absolute, equality of numbers of electors within electoral districts. The point of present significance is the emphasis in the Canadian decisions upon the different historical traditions in the United States and Canada which have shaped the concept or principle of representative government.
It should follow from what I seek to show later in these reasons as to the evolution of representative government in Australia that, in this respect, Australia has far more in common with Canada than the United States.
The adaptation of representative government to federalism by the framers of the Constitution
One point of significance for the present case which does emerge from consideration of the United States decisions is the need to consider the doctrine or principle of representative government not at large but in its adaptation to federalism. In Australia, the framers of the Constitution approached their task with this in mind. They performed their task in such a way as to allow room for further legislative evolution in the system of representative government. Constitutional rigidity was, to a significant degree, avoided. But, in a sense, the plaintiffs seek to assert and to rely upon such rigidity. This is imposed, as their submissions would have it, upon the legislature as a matter of necessary implication from the constitutional text, and is to be lifted only by referendum passed under s 128 of the Constitution.
To reject that approach to the case is not to deny that in significant respects the framers of the federal system did leave much to inference [414] .
Responsible government has been said in this Court to be a central feature of the system devised by the framers of the Australian instrument of government [415] . It appears not from express terms so much as from the requirement in the last paragraph of s 64 that a Minister be a member of the Senate or the House of Representatives. Other central features are the separation of the judicial power of the Commonwealth and the doctrine of judicial review of legislative act and executive decision for constitutional validity. These are essential elements of federalism, but again, to a significant degree, the result is to be seen as a matter of necessary inference from consideration of the text and structure of the Constitution [416] .
The Constitution also established for the Commonwealth, and prescribes and gives effect to, a system of representative government. This was accepted by the whole Court in Australian Capital Television [417] .
To adopt as a norm of constitutional law the conclusion that a constitution embodies a principle or a doctrine of representative democracy or representative government (a more precise and accurate term [418] ) is to adopt a category of indeterminate reference. This will allow from time to time a wide range of variable judgment in interpretation and application [419] . That, of itself, may not be open to objection. However, difficulty can arise where the wide range for variable judgment depends upon, or at least includes as a significant element, matters primarily or significantly of political weight and estimation. It is this concern which, at bottom, founds the celebrated dissenting judgments of Frankfurter J in Baker v Carr [420] and of Harlan J in Reynolds v Sims [421] .
At the heart of the submissions for the plaintiffs is the significance they seek to attach to the entrenched incidents of representative government, both as established by the Constitution for the new federal government and as, in their submission, carried forward by the Constitution, particularly by s 106, as a requirement for the Constitutions of the respective States. The position which the plaintiffs contend is established as regards the Commonwealth thus is the first, and vital, stepping-stone in their argument on this branch of the case.
However, in approaching the matter in this way, it should be kept steadily in mind that there is a fundamental difference involved in comparison between the system of government created for the Commonwealth and that which obtains for the States. As I have said, in the case of the Commonwealth, it was necessary to adapt notions of representative government to the requirements of federalism as hammered out in forming the federal compact. The framers of the Constitution had before them the systems of government operating in the United States and Canada. It should not be overlooked that, at first hand, they had the example of the failed Federal Council of Australasia. This had been established pursuant to the Federal Council of Australasia Act 1885 Imp [422] (the 1885 Act). It suffered from the absence of any directly elected legislative arm. The most obvious impact of federalism upon the federal legislature has been in the establishment and composition of the Senate. However, the States have been free to adapt representative government to bicameral or single legislatures without such federal constraints. These considerations suggest, at the outset, the need for some caution in any automatic translation to the States, through the prism of s 106 of the Constitution, of representative government as established for the Commonwealth.
The Constitution was framed against a background of differing mechanisms for the implementation in the Australian colonies of what in each colony would have been regarded as representative government. The minimum age for candidacy for the upper houses in Victoria, Tasmania and South Australia was thirty years as against twenty-one for all the other chambers. Women were enfranchised only in South Australia and Western Australia and were eligible as candidates only in South Australia. Several States imposed property qualifications upon electors for one or both chambers of the legislature.
From this diversity the challenge was to produce uniformity at the new and federal level. There were two difficult and vital needs. The first was to adapt representative government to the central government of the federation. The second was to achieve this without denying room for further development in the institutions of representative government. These tasks were performed by setting down in the Constitution a variety of prescriptions about parliamentary elections, some of which were entrenched until the Constitution itself might be amended and others which were to prevail only until the Parliament otherwise provided.
The architects of the Australian federation shared an expectation that the federal Parliament would embrace what were then advanced ideas of political representation. The Federal Council of Australasia had not been an elected body. The Canadian Senate (then as now) and the upper houses in New South Wales and Queensland were nominated for life. Nor was the United States Senate then (before the ratification in 1913 of the Seventeenth Amendment) directly elected by the people of the several States. The provisions of ss 7 and 24 of the Constitution providing in the former case that the senators of each State would be "directly chosen by the people of the State" and, in the latter, that members of the House of Representatives would be "directly chosen by the people of the Commonwealth" were thus of contemporary significance.
Learned commentators observing the situation from a vantage point outside Australia wrote of the extremely "democratic" nature of the new Constitution, representing "the high-water mark of popular government" [423] . The Canadian authority, Lefroy, and the British savant, Bryce, noted that the Australian Senate was to be more democratic in its composition than the American Senate which then was elected by the various State legislatures; they regarded as reflecting the same democratic spirit the provision (s 28) that the House of Representatives was to continue only for three years from the first meeting of the House, whereas the Canadian House of Commons continued for five years (British North America Act 1867 Imp, s 50) [424] . Moreover, the provision in Art V for amendment of the United States Constitution, unlike that in s 128 of the Australian Constitution, does not require direct participation of the electors at a referendum.
In 1861, John Stuart Mill published his Considerations on Representative Government in which he restated opinions already put forward in articles and essays. Chapter III was headed "That the Ideally Best Form of Government is Representative Government" and contained the passage [425] :
There is no difficulty in showing that the ideally best form of government is that in which the sovereignty, or supreme controlling power in the last resort, is vested in the entire aggregate of the community; every citizen not only having a voice in the exercise of that ultimate sovereignty, but being, at least occasionally, called on to take an actual part in the government, by the personal discharge of some public function, local or general. Chapter V, titled "Of the Proper Functions of Representative Bodies", commenced [426] : In treating of representative government, it is above all necessary to keep in view the distinction between its idea or essence, and the particular forms in which the idea has been clothed by accidental historical developments, or by the notions current at some particular period.
The meaning of representative government is, that the whole people, or some numerous portion of them, exercise through deputies periodically elected by themselves the ultimate controlling power, which, in every constitution, must reside somewhere. This ultimate power they must possess in all its completeness. They must be masters, whenever they please, of all the operations of government. There is no need that the constitutional law should itself give them this mastery. It does not in the British Constitution. But what it does give practically amounts to this.
Mill continued: But while it is essential to representative government that the practical supremacy in the state should reside in the representatives of the people, it is an open question what actual functions, what precise part in the machinery of government, shall be directly and personally discharged by the representative body. Great varieties in this respect are compatible with the essence of representative government, provided the functions are such as secure to the representative body the control of everything in the last resort. 1. Mill, Utilitarianism, Liberty and Representative Government, Everyman Edition (1957), p 207. 2. Mill, Utilitarianism, Liberty and Representative Government, Everyman Edition (1957), pp 228-229.
As Mill himself recognised, there is room for divergence between particular structures of government, all of which nevertheless might properly be said to constitute representative government. And, of course, those structures might be modified from time to time whilst still providing for representative government. Some incidents of representative government themselves might change from time to time. Any writer upon the British Constitution, a testimony to the forces of evolution, could not have expected otherwise. Against this background and understanding, the expression "representative government" was used by Mill as a "generic description" [427] of a range of political institutions. These had as their essence the placing of ultimate controlling power with the people, to be exercised by representatives of the people elected periodically in free elections to a legislative chamber or to the more powerful chamber of a bicameral legislature [428] .
Mill acknowledged that his broad ideas required refinement in their application to a bicameral legislature [429] . This also is the case with a constitution which is written and rigid rather than fluid, and in particular with a federal system which includes a bicameral legislature established by a constitution containing its own mechanism for change. In Ch 17 of his work, Mill wrote [430] that the fitting constitution of a federal government included a legislative branch, to the composition of which there might be adapted the general principles of representative government. He regarded the provisions of the American Constitution in this respect as "exceedingly judicious", in particular in the representation in the Senate of each State, whether large or small, by the same number of legislators.
The Convention Debates [431] , and the lengthy debates [432] on the Bill for the Commonwealth Electoral Act 1902 Cth (the 1902 Act) manifest a familiarity on the part of significant figures in the federal movement such as Barton, Deakin, O'Connor, Symon, Downer and Glynn with the writings of Mill, Bryce, Thomas Hare and Jethro Brown and, in Australia, the writings of Professor E J Nanson, Catherine Helen Spence and others. Much of this literature concerned the appropriate method of representation in the legislature, the use of single or multi-member constituencies, "first past the post" voting, preferential voting and proportional representation.
Mr Hare's treatise, The Election of Representatives, was published in 1859 and drew the support of Mill in his work on representative government. The implementation of Hare's theories was to await the introduction by Inglis Clark of a Bill in Tasmania in 1896 [433] . The Hare-Clark system, as it came to be called, was adopted (by regulations made under the Federal Elections Act 1900 Tas) for the elections held in March 1901 for the six senators and five members of the House of Representatives.
Mill, Hare and the other writers have been described as idealists who had a continuing influence upon the formulation of electoral laws in Australia [434] . There was concern lest the electoral system fail to afford representation of minority opinion and groups by denying full "value" to the ballot cast by the individual. But there also were pressures of pragmatism. The concern here was with the "value" of the individual vote, not so much to the elector who cast it as to the political party whose candidate it favoured, and fear of "weak government" flowing from a legislature in which there were more than two principal parties represented.
One concludes that at the time of federation issues of "representative government" which concerned the franchise and the electoral system involved questions of political philosophy, of degree, and of experience of practical affairs. Senator Sir Josiah Symon said that John Stuart Mill had supported a great many things which were theoretical and that he should be sorry to accept any of the views of Mill and his supporters "in the face of the opinions of practical politicians" [435] .
Given the special adaption of principles of representative government to federalism, where in such a case does ultimate sovereignty reside? Writing in 1901, Bryce put the view that, in such a case, ultimate sovereignty resides with the authority or body which, according to the constitution, may amend the constitution [436] .
In Australia, that ultimate authority, to change the Constitution itself, is reposed by s 128 in a combination of a majority of all the electors and a majority of the electors in a majority of States. But the initiative to place before the electors any proposed change is vested in the Parliament and a particular mechanism is provided in s 128 to resolve disagreement between the Houses as to the passage of a proposed law for the alteration of the Constitution. Special provision is made in the final paragraph of s 128 in respect of proposed alterations (i) diminishing the proportionate representation of any State in either House, or the minimum number of representatives of a State in the House of Representatives, (ii) altering the limits of a State, or (iii) "in any manner" affecting the provisions of the Constitution in relation to a State. In such cases, the proposed law effecting the change is not to become law unless approved by the majority of electors voting in the State concerned.
It is to be observed that, in significant respects, s 128 does not provide for an equality of voting power at referendums. A negative power, in other words a power to reject changes, may be exercised by a minority of the total electors of the Commonwealth if that minority is geographically distributed such as to constitute a majority in a majority of States. Again, the electors in a particular State must approve a proposed law affecting that State in the ways outlined above. Thus, a majority of electors in that one State may resist a change otherwise approved by a majority of electors in a majority of States. In both cases the value of each elector's vote is not equal in the sense for which the plaintiffs contend.
Broad statements as to the reposition of "sovereignty" in "the people" of Australia, if they are to be given legal rather than popular or political meaning, must be understood in the light of the federal considerations contained in s 128.
Those statements must also allow for the fact that none of the Australia Acts, Imperial, Commonwealth or State, followed approval at a referendum, in particular, any submission to the electors pursuant to s 128 of the Constitution. Moreover, in s 15 thereof, the Australia Acts provide their own mechanism for amendment or repeal by statute and without submission to the electors at State or Commonwealth level.
In certain respects, the Constitution was drawn so as to give mandatory effect to some elements of representative government. But this was done to answer the perceived needs of a federal structure of government. Section 24 of the Constitution reveals four elements of representative government which were perceived as "essential" to the federal structure - at least in the sense that to change them would require constitutional amendment [437] .
(1) "The House of Representatives shall be composed of members directly chosen by the people of the Commonwealth."
In McKinlay [438] , Gibbs J said that the "obvious purpose" of these words was to ensure that members of the House not be chosen "by some indirect means, such as by the Parliament or Executive Government of a State, or by an electoral college". In the same case, it was said by some members of the majority that this phrase was emphatic of a further factor, that there be a popular election or popular vote [439] , so that a sufficient inequality of distribution of numbers between electoral divisions might mean that there was no choice by the people of the Commonwealth [440] . As I have indicated, the plaintiffs rely upon what was said on this point in McKinlay , and it will be necessary to return to their submissions.
(2) "[T]he number of such members shall be, as nearly as practicable, twice the number of the senators."
The Constitution pegs the number of members of the House of Representatives to the number of senators. Within that limitation the Constitution gives to the Parliament the power to make laws increasing or diminishing the number of senators for each State (s 7), and also the power to make laws for increasing or diminishing the number of members of the House of Representatives (s 27).
(3) "The number of members chosen in the several States shall be in proportion to the respective numbers of their people."
The concern of the framers here was to ensure some equality of voting power by requiring proportionality between the population of each State and the number of members which that State could elect. In other words, as between the populations of each State, there was to be an equality of voting power because the ratio of the number of members chosen in each State to the total number of members was to be the same as the ratio of the population of that State to the total population.
(4) "[F]ive members at least shall be chosen in each Original State."
The framers entrenched a requirement that the electors of each Original State, however small, should choose at least five members for the House of Representatives. The point may be made here that this effectively denied the possibility of "one vote, one value" at federation by arbitrarily boosting the value of votes in the smallest States [441] . (It may also be noted that the second paragraph of s 7 authorises the Queensland Parliament, until the federal Parliament otherwise provides, to divide Queensland into divisions and to determine the number of senators for each division [442] ).
These elements were thought by a majority of delegates to be sufficiently important to justify entrenchment as constitutional requirements. The entrenched matters concerned issues of federalism.
The pegging of the number of members of the House of Representatives to the number of senators was designed to preserve the integrity of the Senate in its relations with the House of Representatives in at least two senses. First, the ratio of members of the House to senators is of crucial significance in the event of any disagreement between the chambers and the engagement of the process laid down in s 57 to resolve such disagreement. Second, there was a perception amongst the delegates that any greater disparity between the number of members of the House of Representatives and senators was inimical to the prestige and standing of the Senate [443] .
The requirement in s 24 that the number of members chosen for each State shall be in proportion to the respective number of its people is evidence of a concern for equality of voting power, but only in the limited sense of interstate equality. This was purely a matter of apportioning the number of members which were to be chosen by the electors of each State. There was no suggestion that, within each State, members were to be chosen from electoral divisions of equal numbers of electors. The concern was one of equality of voting power among States, not among people.
Some further insight is to be gained from a recognition that even those elements, which I have identified as having been deemed sufficiently important to justify constitutional entrenchment, were not universally accepted at the Conventions. There was concerted opposition at Adelaide to the constitutional requirement of a 2:1 ratio between members of the House of Representatives and senators [444] . Opposition to the requirement [445] was based upon a preference for an alternative method of calculating the number of members for the House of Representatives which were to be chosen by the electors of each State. Sir George Turner moved that the requirement of a 2:1 ratio be struck out and replaced with the words "and until the Parliament of the Commonwealth otherwise provides each State shall have one representative for every 50,000 of its people" [446] . This marked a return to the form of the requirement which had passed the Convention in 1891 [447] . It is to be noted that this was only another requirement as to numerical apportionment of the number of members which were to be elected by the population of any one State. It was not a requirement that electoral divisions were to be such that each had 50,000 electors, thus ensuring, on one measure, equality of individual voting power. The issue was one of federalism, not equality.
Some opposition to the constitutional entrenchment of the four matters which I have mentioned was based upon the view that these matters, along with the rest of Pt III of Ch I, were better left to Parliament. For example, in 1897, Mr Deakin said [448] :
We cannot possibly foresee the future, whether as to the gathering together or the dispersal of population in these colonies, or the division of colonies as they at present exist, nor define where the great and perhaps overwhelming accretions of population will be. Under these circumstances, why lay down an iron rule for changing conditions that will involve great expenditure and great difficulty in order that it may be altered. [sic] We are, as it seems to me, justified, and well justified, in embodying in this Constitution general principles which will relate to the permanent conditions of the Commonwealth, those which will obtain for all time, but this is distinctly a changing condition. 1. Convention Debates, vol 3, pp 701-702.
What does appear in ss 7 and 24 of the Constitution is the phrase "directly chosen by the people". However, upon examination, this does not have the general significance sought to be drawn from it by the plaintiffs.
At federation, the upper houses in the colonies of New South Wales and Queensland were nominated for life. Further, the framers of the Constitution had before them the spectre of the failed Federal Council of Australasia. This was despatched into history by covering cl 7 of the Constitution. Sections 5 and 6 of the 1885 Act, repealed by covering cl 7, had provided for each participating colony to be represented by two members to be appointed under provision made by the legislature of that colony. Queensland, Victoria, Tasmania and South Australia provided [449] for appointment by the Governor from members of either chamber of the legislature. In Queensland, where the upper house was a nominated body, representatives might also be selected from those who had been members of either house at some time within the last six months. In Queensland and South Australia, members might also be appointed from the Executive Council, and in Western Australia [450] , which did not then have responsible government, appointment was by the Governor from an unlimited field.
This contemporary background of a perceived deficiency in the structure which the Constitution replaced provides strong support for the conclusion expressed by Gibbs J in McKinlay [451] that the obvious purpose of the opening words of s 24 (as well as the effect of the text) was to deny any method of indirect election by requiring direct selection by popular vote. The same, in my view, is true of s 7, as regards the Senate. In each case the provision strengthened the new federal system by requiring that neither chamber be composed of members selected by or from any legislative chamber of a State or by the executive government of a State or (a remote possibility, given the unsatisfactory experience in the United States) by an interposed electoral college.
The phrase in s 24 "directly chosen by the people of the Commonwealth" is a broad expression to identify the requirement of a popular vote. Section 41 (considered later in these reasons) operated to secure in the federal franchise the female franchise which had been acquired in South Australia and Western Australia. In other respects, the selection of those from among the population who were to be empowered to make the electoral choice was left by s 24 to s 30 of the Constitution. Subject to denial of plural voting, this left the matter at large for provision by the new Parliament. In the meantime, qualification was left to the miscellany of laws in the several States which dealt with elections for the lower houses. Section 29 provided for determination of electoral divisions by the Parliament in substitution for State laws upon the matter.
It follows that, in my view and contrary to that of Murphy J in McKinlay [452] , the phrase in s 24 is not to be dissected in the way the plaintiffs would have it so as to provide a distinct component "chosen by the people", with its own operation above that of s 24 as a whole and independently of ss 29 and 30. It also follows that McKinlay necessarily still stands for the somewhat narrower, or at least discrete, proposition, also negative in nature, that s 24 does not require equality in electoral divisions.
To reject this branch of the plaintiffs' case is not necessarily to reject their broader submission that malapportionment, if sufficiently serious, may deny the system of representative government established for the Commonwealth by the Constitution, taken as a whole. Before further considering that submission, it is appropriate to consider the allowance for the evolutionary nature of representative government made by the Constitution.
In many respects, the architects of the Constitution "placed great faith in the capacity of the elected senators and members to design statute law for a system of representative self-government, notwithstanding that they would be legislating in their own interest" [453] . But, as a practical matter, it had been necessary to rely upon later provision by the Parliament because, upon many matters concerned with the system of representative government, the framers of the Constitution had been unable to achieve agreement. O'Connor [454] expressed the reality of the situation when, after referring to the "provisional condition of things" contemplated by various provisions of Ch I of the Constitution, he said:
When we sat in Convention it would no doubt have been more in accordance with the wishes of the delegates, and more in accordance with the proper working of the Constitution, if we could at that time have arrived at some uniform method by which the people of Australia could be represented in the Parliament of Australia. But it was obviously impossible at that time to arrive at any system of the kind; and, therefore, in order that a workable basis for the first Parliament might be created, we adopted the different franchises and methods of the different States and declared that, until the Parliament provided, that should be the working basis of our Constitution. 1. Reid and Forrest, Australia's Commonwealth Parliament 1901-1988, (1989), p 87. 2. Speaking as the Minister with the carriage in the Senate of the Bill for the Commonwealth Electoral Act 1902, Parliamentary Debates (Hansard), 1902, vol 8, p 10702.
The point may be illustrated by an example germane to the present litigation. In McKinlay [455] , Mason J said that it was significant that the second paragraph of s 24 of the Constitution made explicit provision for the ascertainment of a quota of people for the specific yet limited purpose of determining the number of members chosen in each State and pointed out that, had it been intended to provide that electoral divisions within a State should contain an equal or practically equal number of people or electors, it was "inconceivable that the quota provision would have been expressed to have such a limited application". In the draft adopted at the 1891 Sydney Convention, cl 24 provided that "until the Parliament of the Commonwealth otherwise provides, each State shall have one Representative for every thirty thousand of its people". That provision did not reappear in the draft brought up to the 1897 Convention and despite debate upon the matter was not reinstated [456] .
Nevertheless, the recurrent phrase "until the Parliament otherwise provides" has a deeper significance. Its effect is to accommodate the notion that representative government is a dynamic rather than a static institution and one that has developed in the course of this century. The accommodation is effected in the Constitution itself by authorising the legislature to make appropriate provision from time to time. It is by this means that the Constitution continues to speak to the present and allows for development of the institution of government by changes which may not have been foreseen a century ago [457] or, if foreseen by some, were not then acceptable generally.
The phrase "until the Parliament otherwise provides" (thereby engaging the head of legislative power in s 51(xxxvi)) appears throughout Pts II, III and IV of Ch I of the Constitution. In respect of the Senate, it appears in s 7 (the people of the State to vote as one electorate until the Parliament otherwise provides and there to be six senators for each original State until the Parliament otherwise provides), s 10 (application to Senate elections of State laws until the Parliament otherwise provides), s 22 (one-third of the whole number of senators to be a quorum until the Parliament otherwise provides). As regards the House of Representatives, the phrase appears in s 24 (determination by a particular method of the number of members chosen in the several States until the Parliament otherwise provides), s 29 (electoral divisions), s 30 (qualification of electors), s 31 (application of State laws to elections for the House), s 34 (qualifications of members), s 39 (quorum in the House of Representatives). Sections 46, 47 and 48 apply to both houses and specify, "[u]ntil the Parliament otherwise provides", particular provisions dealing with penalties for sitting when disqualified, disputed elections and allowances to members. Further, the Parliament was empowered by s 49 to declare the powers, privileges and immunities of the Senate and the House of Representatives and (s 27) the Parliament was authorised to make laws increasing or diminishing the number of members of the House of Representatives. Section 51(xxxvi) empowers the Parliament to make laws with respect to matters in respect of which the Constitution makes provision until the Parliament otherwise provides.
What is of present significance is that, in some provisions of Ch I of the Constitution, express choices were made as to particular incidents of that system of representative government which was to be established for the Commonwealth. I have referred already to the provisions in ss 7 and 24 that each House was to be composed of those "directly chosen by the people" of the State in the case of the Senate, and of the Commonwealth in the case of the House of Representatives. Further, in choosing members of the Parliament, it was ordained that "each elector shall vote only once" (ss 8 and 30). Whilst plurality of voting might today seem anachronistic, it should be noted that it existed at the time of the adoption of the Constitution in the franchises of Tasmania, Queensland and Western Australia [458] and that it lingered in Britain until removed by s 1(2) of the Representation of the People Act 1948 UK [459] (the 1948 Act).
Section 43 rendered a member of one chamber ineligible to be chosen for or to sit in the other chamber; this was founded on the constitutional practice of the Imperial Parliament. The qualifications for a senator were to be the same as those for a member of the House of Representatives (s 16). The disqualification provisions of s 44 rendered certain persons incapable of being chosen or of sitting either as a senator or as a member of the House of Representatives. The method of choosing senators was to be uniform for all the States, whether prescribed by a law of the Commonwealth or by State law (s 9).
An electoral division was not to be formed out of parts of different States (s 29). However, s 29 contemplated laws which would provide for more than one member to be chosen for a division. As late as 1946, in the British House of Commons, 18 constituencies returned two members and one returned three members [460] . Single member constituencies were made necessary in Britain by s 1(1) of the 1948 Act. Section 12 of the 1902 Act provided that one member of the House of Representatives should be chosen in each division. Multi-member constituencies for the lower houses existed in 1900 in Victoria, Queensland, South Australia (where all but two of 27 divisions returned two members) and Tasmania (where Hobart returned six members to a chamber of 30). In the debate on the Bill for the 1902 Act, Senator Sir John Downer and his supporters in the Senate favoured both multi-member constituencies and voting with proportional representation as means of avoiding the tyranny by bare majority and of advancing the interests of those voters who were in the minority [461] . Kingston's Draft Constitution prepared for the 1891 Convention in Sydney had provided in Pt VII for a National Assembly with each district to return two members.
The existence of female franchise in South Australia and Western Australia was, at the federal level, protected by s 41 of the Constitution. The effect of s 41 was that no adult person who, before the establishment of the federal franchise, had acquired a right to vote at a State election was to be prevented by any law of the Commonwealth from voting at elections for either house of the federal Parliament [462] . Further, until the qualification of electors of members of the House of Representatives became uniform throughout the Commonwealth (by reason of the adult franchise conferred by s 3 of the Commonwealth Franchise Act 1902 Cth), s 128 provided that in any State with adult suffrage only one-half of electors voting for or against a referendum were to be counted.
However, the Constitution did not entrench the secret ballot, compulsory voting, preferential or proportional voting, nor any universal adult franchise. Nor did the Constitution prescribe any authority or body to determine from time to time the electoral divisions in each State.
Compulsory enrolment for federal elections and for referendums was introduced by s 8 of the Commonwealth Electoral Act 1911 Cth and compulsory voting at referendums was introduced by the Compulsory Voting Act 1915 Cth. Compulsory voting in elections was introduced by s 2 of the Commonwealth Electoral Act 1924 Cth and the validity of that law was upheld in Judd v McKeon [463] . Professor Crisp wrote that both major parties were in favour of the 1924 legislation though neither wished to take overt responsibility for it; hence it was introduced as a private member's bill on which no Minister spoke and no division was called in either House [464] . The Commonwealth Electoral Act 1918 provided (s 124) for the introduction of preferential voting for House of Representatives elections and this was achieved for the Senate by s 7 of the Commonwealth Electoral Act 1919 Cth. Finally, proportional representation in the Senate was introduced by s 3 of the Commonwealth Electoral Act 1948 Cth [465] .
There is considerable force in the recent statement by learned commentators [466] :
As numerous and as positive in expression as many of these [constitutional] provisions are, they constituted only the bare foundations of the electoral law for the representative Parliament of a new nation. The Constitution, for example, left unspecified, or open to change, a whole range of matters including: the method of voting to elect the members of the respective houses; the question of whether members of the House of Representatives would be elected by single-member or multi-member divisions; the length of time each State would continue to vote as one electorate in electing the Senate; who would be authorised to vote; the question of voluntary or compulsory registration of voters and of voting itself; the control of electoral rolls; the conduct of the ballot; the style of ballot papers; the use of postal votes; limitations on the electoral expenses of candidates; the financial deposits to be made by candidates and the conditions of their forfeiture; the role of political parties at elections; the question of financial support for political parties from public funds; the location of responsibility for the administration of the electoral law; and the extent of the delegation of authority in electoral decision-making. 1. Reid and Forrest, Australia's Commonwealth Parliament 1901-1988 (1989), pp 86-87.
To conclude from an examination of the express provisions of the Commonwealth Constitution that it, or the Constitutions of the States, enshrines or at least adopts some of the principles of representative government is not of any immediate assistance in resolving the specific legal issue with which this case is concerned. This is the alleged invalidity of the legislation pursuant to which the State of Western Australia is divided geographically into districts and regions in such a manner as is said not to give "equal value" to the vote of each elector, whether for the Legislative Assembly or the Legislative Council.
This notion of "equal value" was not, in my view, an essential or inherent feature of the system of "representative government" as understood at federation. It may properly be said (and this Court has held it to be the case) that the Constitution established and continues to prescribe such a system for the Commonwealth (as does the Western Australian Constitution for that State).
The particular "entrenched" provisions of the Constitution to which I have referred gave effect to the essential character of representative government which Mill had identified. But, as McHugh J pointed out in Theophanous [467] , the Constitution did not specify "the whole apparatus of representative government". As to much of that, it was, as Barton had said in 1891, a case of "trust the parliament of the commonwealth" [468] . The Constitution explicitly proceeds on that footing.
The Australian colonies in 1900 had determined, by various means, the geographical areas and numbers of persons or of electors in the electoral divisions, and none insisted on practical equality in numbers of people or electors in those divisions [469] . The materials in the present case show considerable margins of deviation in those electorates. At that time, in the United Kingdom, there remained substantial inequality between electorates, as Gibbs J pointed out in McKinlay [470] . The legislation in force in New South Wales in 1900 [471] permitted electorates to be larger or smaller than the average quota by a considerable margin. In Victoria in 1901 there were electorates of about 9,000 electors and at least one electorate of 16,000 [472] . Western Australia, as I have indicated earlier in these reasons, presented a stark picture of great extremes.
At the federal level, s 16 of the 1902 Act provided that the quota of electors represented by the division of the whole number of electors in the State by the number of members of the House of Representatives to be chosen in the State was not to be departed from to a greater extent than one-fifth more or less. It should also be noted that the criteria adopted in s 16 of the 1902 Act for the drawing by the Commissioner of boundaries of electoral divisions for the House of Representatives included physical features, means of communication and community or diversity of interest. Thus, from the outset, federal electoral law recognised the point later made with some force by Harlan J in Reynolds v Sims [473] :
[P]eople are not ciphers and legislators can represent their electors only by speaking for their interests - economic, social, political - many of which do reflect the place where the electors live [C]onflicting interests within a State [cannot] be adjusted by disregarding them when voters are grouped for purposes of representation. 1. (1964) 377 US 533 at 623-624.
It does not follow from the prescription by the Constitution of a system of representative government that a voting system with a particular characteristic or operation is required by the Constitution. What is necessary is the broadly identified requirement of ultimate control by the people, exercised by representatives who are elected periodically. Elements of the system of government which were consistent with, albeit not essential for, representative government might have been constitutionally entrenched or left by the Constitution itself to the legislature to provide and modify from time to time. This is what was done.
Nor if, contrary to what was decided in Leeth v The Commonwealth [474] , there is a doctrine of the underlying equality of the people of the Commonwealth before the law, would it apply so as to require a different outcome.
Looked at in this way, the task of the plaintiffs then becomes to demonstrate that the voting system of which they complain is so distorted as not to answer the broad identification (which they contend is implicit in the system of representative government established by the Constitution for the Commonwealth and carried by force of the Constitution into the Constitution of the State) of ultimate control by periodic popular election.
To analyse the issues in this case in such a manner is not to deny the important proposition established, as regards the Commonwealth, by Australian Capital Television [475] . This is that legislation otherwise within power will be invalid if it attacks the structure of the government of the Commonwealth by infringing in the requisite sense freedom of communication on matters relevant to political discussion, that freedom being an inherent characteristic of the system of representative government recognised and provided for in the Constitution. Although this principle is capable of limiting the exercise of the legislative powers of the State Parliaments, it is concerned with protection of the governmental structure of the Commonwealth [476] .
The range of provisions in the Constitution which provide for particular incidents of representative government "until the Parliament otherwise provides" and the provision in s 128 for amendment of the Constitution by referendum approved by the electors emphasise the central importance of communication on matters relevant to the efficacious working of the system of representative government established for the Commonwealth. It is hardly to be expected that the Constitution was framed so as to present an impermanent or incomplete statement of the incidents of responsible government on the footing that the Parliament which would make changes and remedy deficiencies perceived from time to time would be composed other than by the representatives of electors who had been free of legislative impediment in informing themselves and in receiving information and comment upon matters of political interest.
I do not regard Australian Capital Television as authority for any wider or more general propositions or principles.
In this litigation, no challenge is made, and I am far from suggesting that any well-based challenge might be made, to the system of electoral divisions, established by Pt IV (ss 55-78) of the 1918 Act, on the footing that there has been a departure from the broad constitutional requirement of ultimate control of the composition of the House of Representatives by the people of the Commonwealth who exercise their control through the medium of popular elections.
I would accept that the variations in numbers of electors or people in single-member divisions could be so grossly disproportionate as to deny ultimate control by popular election [477] . I would, with respect, also agree with the point made by McTiernan and Jacobs JJ in McKinlay [478] that, when it arises, such a question is to be determined by reference to the particular stage which then has been reached in the evolution of representative government. By way of example, for both Commonwealth and State elections, provision has been made for more than 20 years for 18 as the minimum age for voters [479] . An even plainer example is the now long-established universal adult suffrage. This has become a characteristic of popular election of senators and members of the House of Representatives which could not be abrogated by reversion to the system which operated in one or more colonies at the time of federation. In my opinion, this is so notwithstanding that ss 8 and 30 of the Constitution, subject to the prevention of plural voting, permitted the qualification of electors to be ascertained in that way, until the federal Parliament otherwise provided.
I agree also with the statement by McTiernan and Jacobs JJ in the same passage in McKinlay to the effect that the point at which there ceases to be a system of representative government because there is a failure in ultimate control by periodic popular election involves a question of degree and is one which cannot be determined in the abstract. A particular reason why this is so is provided in Dixon v Attorney-General (British Columbia) [480] . As I have indicated, McLachlin J was there applying the express provision in s 3 of the Canadian Charter of Rights and Freedoms. Nevertheless, the reasoning by which the Court determined that the British Columbia legislation establishing provincial electoral districts violated s 3 involved steps of present significance.
The first consideration is that the historical development of voting rights in Canada (and I would consider the same was true in Australia) leads to the conclusion that relative equality of voting power may be seen as the dominant principle underlying the system of representative government as it has evolved, so that relative equality in numbers of electors is the single most important factor in determining electoral boundaries [481] . The second, developed in McLachlin J's later judgment in the Supreme Court of Canada in Reference re Electoral Boundaries Commission Act [482] , is that "representation" comprehends not only a voice in the deliberations of government but also the right to bring one's grievances and concerns to the attention of a local member of the legislature. The third is that relative parity in numbers of electors in electoral divisions may detract from effective representation; to repeat what was said in Reference re Electoral Boundaries Commission Act [483] :
Factors like geography, community history, community interests and minority representation may need to be taken into account to ensure that our legislative assemblies effectively represent the diversity of our social mosaic. These are but examples of considerations which may justify departure from absolute voter parity in the pursuit of more effective representation; the list is not closed. Finally, whilst relative parity in numbers remains the single most important concern, determination of whether the legislation in question has, in a given case, operated to deny effective representation in the sense described above, requires recognition by the judicial branch of government that the legislative branch is in a better position to weigh, or to establish specialised administrative bodies to weigh, the various factors which are involved. 1. Dixon (1989) 59 DLR(4th) 247 at 266. 2. [1991] 2 SCR 158 at 183; (1991) 81 DLR (4th) 16 at 35. 3. [1991] 2 SCR 158 at 184; (1991) 81 DLR (4th) 16 at 36.
In that connection, I turn again to the judgment of McLachlin J in Reference re Electoral Boundaries Commission Act [484] where her Ladyship said:
I turn finally to the admonition that courts must be sensitive to practical considerations in interpreting Charter rights. The "practical living fact", to borrow Frankfurter J's phrase [in Baker v Carr [485] ], is that effective representation and good government in this country compel those charged with setting electoral boundaries sometimes to take into account factors other than voter parity, such as geography and community interests. The problems of representing vast, sparsely populated territories, for example, may dictate somewhat lower voter populations in these districts; to insist on voter parity might deprive citizens with distinct interests of an effective voice in the legislative process as well as of effective assistance from their representatives in their "ombudsman" role. This is only one of a number of factors which may necessitate deviation from the "one person - one vote" rule in the interests of effective representation. 1. [1991] 2 SCR 158 at 187-188; (1991) 81 DLR (4th) 16 at 38. 2. (1962) 369 US 186.
If the particular legislation under challenge establishes a system which requires a specialised body or tribunal to define electoral divisions in a manner which gives effect to these principles, then the legislation itself will not be open to challenge as denying the constitutional requirement for representative government of ultimate control by popular election. Rather, in any given case, it will be a question to be determined by reference to the circumstances of that case whether judicial review is required to ensure that the exercise in the particular case of the authority conferred by the legislature is confined within constitutional limits [486] .
That being the position as regards the Commonwealth, the question then is whether, by force of the Constitution, any (and if so what) restraint is imposed upon the legislative power of the States to provide for the manner of representation in their legislatures.
In my view, it is not an acceptable mode of argument to take the present electoral system in Western Australia and contend that it offends some restraint because, if it applied to elections for the House of Representatives (and, semble, the Senate), it would involve such gross disproportion between constituencies as to deny ultimate control by popular election and so entail legislative invalidity. Geographic and demographic factors apply at the national level and the levels of the several States with significant differences in scale and quality. Like is not being compared with like.
In addition, I have already referred to the caution required in translating to the State sphere any doctrine which springs from the adaption of representative government to the legislative branch of the government of the federation. I have also referred to the United States experience which is that the nature of the interests to be represented at State level may be more diverse and more parochial in nature than at the national level. That also, in my view, is true of Australia.
Further, no relevant restraint of the kind contended for by the plaintiffs applies by force of the Constitution to the legislatures of the States in the exercise of their power to provide for methods of voting and apportionment of electors between electoral districts or regions. In dealing further with this aspect of the case, it is necessary to have regard to a number of recent decisions of this Court in order to determine their immediate significance.
In Stephens [487] a majority of the Court expressly accepted that, from the text of the 1889 Act as it now stands, and the system of government for which it provides, there is to be drawn an implication of freedom of communication to discuss "government, governmental institutions and political matters" as Brennan J put it [488] . However, as will appear, that was not the only point on which the case turned.
In Theophanous [489] , the majority of the Court proceeded on the footing (which appears to have been common ground to both the parties and the interveners) that, in the earlier decisions of Nationwide News [490] and Australian Capital Television [491] , this Court had "distilled from the provisions and structure of the Constitution, particularly from the concept of representative government which is enshrined in the Constitution, an implication of freedom of communication" [492] or "of political communication and discussion" [493] . The immediate issue in Theophanous was the applicability of that implication in the Constitution upon the common law and statute law of a State which imposed or defined civil liability for defamation. At the time of the publication in question, which criticised the plaintiff's views on immigration (a matter of federal legislative concern), he was a Member of the House of Representatives and, with other offices, held the chair of the Immigration Committee of the federal Caucus of the Australian Labor Party. Thus, there was no issue for decision in that case which required any conclusion as to extension of a federal constitutional implication to political communication and discussion at the State level.
However, in Stephens [494] the plaintiffs were members of the Legislative Council of Western Australia and sued for defamation in respect of publication of assertions by another Council member that their overseas study trip was "a rort which is costing taxpayers thousands of dollars". The case did not turn solely upon an implication of freedom of communication derived from the State Constitution. Deane J [495] accepted, for the purposes of that case, the views of the other members comprising the majority (Mason CJ, Toohey and Gaudron JJ). Their Honours [496] considered that "the freedom of communication implied in the Commonwealth Constitution extends to public discussion of the performance, conduct and fitness for office of members of a State legislature". Their Honours referred to judgments in Nationwide News [497] and in Australian Capital Television [498] and to observations in their joint judgment in Theophanous . In that case [499] , their Honours observed:
The interrelationship of Commonwealth and State powers and the interaction between the various tiers of government in Australia, the constant flow of political information, ideas and debate across the tiers of government and the absence of any limit capable of definition to the range of matters that may be relevant to debate in the Commonwealth Parliament and to its workings make unrealistic any attempt to confine the freedom to matters relating to the Commonwealth government [500] . That said, the question is of little importance in the present case. The publication complained of relates to the views, performance and capacity of the plaintiff as a member of the Commonwealth Parliament. 1. (1994) 182 CLR 211. 2. Stephens (1994) 182 CLR 211 at 257 3. Stephens (1994) 182 CLR 211 at 232 4. (1992) 177 CLR 1 at 75-76, per Deane and Toohey JJ. 5. (1992) 177 CLR 106 at 142, per Mason CJ; at 168-169, per Deane and Toohey JJ; at 216-217, per Gaudron J. 6. Theophanous (1994) 182 CLR 104 at 122 7. Australian Capital Television (1992) 177 CLR 106 at 142, 168-169, 215-217.
Thus, various factors, taken together, and involving aspects of contemporary life and experience such as the constant flow of political information, ideas and debate, made it unrealistic or impracticable (but, semble, not impossible) to confine the implied freedom of political discussion to matters relating to the government of the Commonwealth; the result was to require extension of the constitutional principle to political discussion of matters relating to the government of the States.
Three points should be made. The first is that, as McHugh J explains in his judgment in the present case, the process of constitutional interpretation by which this principle was derived (being an implication at a secondary level), and the nature of the implication (which restrains not only the exercise of legislative, executive or judicial power but also what otherwise would be the operation of the general law upon private rights and obligations) departed from previously accepted methods of constitutional interpretation. If it now were sought to apply the principle then the need for further examination of it would arise.
However, and this is the second point, the present case does not involve any such principle of freedom of political discussion, whether derived from the Commonwealth or State Constitution. Rather, as I have indicated, it is concerned with the essential characteristics of the constitutionally mandated system of representative government, albeit that it is in aid of this that the principle of freedom of political communication operates.
The third point is that there is nothing in the federal constitutional structure which renders it unrealistic or impractical to accept that, compatibly with the federal structure, from time to time the doctrine of representative government may have an adaptation to one or more of the States which does not correspond to that for the Commonwealth.
I have referred to various matters which indicate that this is so. They perhaps begin with the position of the Senate as the "States' House" in the constitutional design, and various aspects of s 24 of the Constitution, including the reservation to Tasmania and Western Australia of at least five members of the House of Representatives at federation.
The Constitution was adopted at a time when, in two of the colonies, the upper houses of the legislature were nominated for life. It may also be recalled that, as Quick and Garran put it [501] :
Queensland was tripartite in interest, the North and the Centre looked forward to Federation, not only for its own sake, but also as a step towards subdivision; whilst Brisbane and the South feared that their trade would suffer from open competition with New South Wales and its metropolis. If the project of dividing Queensland had ever reached fruition, there would seem to have been nothing in the Constitution to prevent, for example, the adaption of the upper house of the Queensland legislature to a body representing the three regional divisions of the State [502] . I have already referred to the accommodation sought to be made in the second paragraph of s 7 of the Constitution, dealing with Senate representation of Queensland, to the internal separation of Queensland into divisions. 1. Annotated Constitution of the Australian Commonwealth (1901), p 162. 2. Such a proposal had been made in 1892 in the Queensland Constitution Bill introduced in the Legislative Assembly by Sir Samuel Griffith but defeated in the Legislative Council. See Joyce, Samuel Walker Griffith (1984), pp 147, 173-175, 191; Legislative Assembly, Parliamentary Debates (Hansard), 26 July 1892, vol 68, pp 792-795; Legislative Council, 25 and 26 October 1892, vol 66, pp 162-167, 169-177.
As it presently stands, the Constitution, in s 15, recognises, as has become the fact in Queensland, that a State legislature may comprise one chamber only. Thus the first sentence of s 15 now states:
If the place of a senator becomes vacant before the expiration of his term of service, the Houses of Parliament of the State for which he was chosen, sitting and voting together, or, if there is only one House of that Parliament, that House, shall choose a person to hold the place until the expiration of the term. But at no stage has the Constitution contained an implication denying to a State the power to change to a unicameral system, if such power otherwise existed [503] . 1. Clayton v Heffron (1960) 105 CLR 214 at 248-249
On the other hand, the Constitution assumes the continued existence of the office of Governor of a State or other chief executive officer or administrator of the State (ss 12, 15 and 110) and the existence of "the Executive Government" of a State (s 119). Further, Ch III of the Constitution brings the State court system within the ambit of the judicial power of the Commonwealth. It assumes the continued existence of State Supreme Courts and the delivery of judgments, making of orders and decrees, and imposition of sentences, of such a nature as in all instances (subject only to regulation prescribed by the federal Parliament) to be appropriate subject matter for the exercise by this Court of the judicial power of the Commonwealth in its appellate jurisdiction under s 73 of the Constitution. Section 77 presupposes the existence from time to time and the operation of a system of State courts in original States and in those later admitted into or established by the Commonwealth, the jurisdiction and powers of which under State law are compatible with the exercise of the judicial power of the Commonwealth upon investment by the Parliament with federal jurisdiction with respect to matters of the description in ss 75 and 76 of the Constitution.
Section 107 of the Constitution identifies the powers of what had been the Parliaments of the colonies as they stood at the establishment of the Commonwealth and provided that they would continue as the powers of the Parliaments of the States unless by the Constitution exclusively vested in the Parliament of the Commonwealth or withdrawn from the Parliament of the State. Section 108 deals with the saving of laws in force in the colonies before they became States. The Constitution operates, in various provisions (ss 7, 9, 29, 123, 124), upon the footing that each State, whether an original State or a State formed under s 124, would have a Parliament and accepts that those legislatures may make laws or other decisions (for example, to fill a casual vacancy in the Senate under s 15 or to consent to alteration of limits) which have an immediate effect upon the operation of the political and territorial composition of the Commonwealth and the composition of the federal legislature. Moreover, pars (xxxvii) and (xxxviii) of s 51 empower the Parliament to legislate upon reference or request, or with the concurrence of State Parliaments.
Accordingly, it may fairly be said that the framers of the Constitution accepted the structure of the governments of the colonies as they stood at federation and, in so far as the doctrines of representative government and responsible government were then understood, they were treated as operating in the new States.
Nevertheless, in my view, and whatever otherwise might be the scope of s 106 of the Constitution (as to which it is unnecessary to express any view), there is nothing in the Constitution to bind the States to any particular subsequent stage of evolution in the system of representative government. In the events that have happened, there has been in the several States development at various paces in that component of representative government which is concerned with the importance of relative parity in the distribution of electors between the constituencies from which members of the legislature are selected. No doubt further changes in that process of evolution may be expected from time to time. But, in my view, that has not been and will not be in obedience to any imperative imposed upon State legislatures by the Constitution.
Finally upon this branch of the case it may be noted that, as regards the Commonwealth Parliament, such evolutionary change is contemplated, as discussed earlier in these reasons, by the phrase "until the Parliament otherwise provides". There is thus no entrenchment of a particular method of distribution of electors between electoral districts. The implication which the plaintiffs seek to draw from the Constitution presumably would apply unless varied or abrogated by the steps required by s 128 of the Constitution. Thus, the State would be denied the elasticity specifically allowed the Commonwealth itself.
There remains for decision the question whether the submissions of the plaintiffs are to be supported on considerations flowing purely from the constitutional arrangements which exist in the State of Western Australia.
In Stephens [504] , Mason CJ, Toohey and Gaudron JJ observed that the legislature of Western Australia consists of the Sovereign, the Legislative Council and the Legislative Assembly, that both chambers are popularly elected (the members of the Council being elected for electoral regions), and that the electors of members of the Assembly are also the electors of members of the Council. Their Honours then summarised as follows the effect of s 73 of the 1889 Act [505] :
Section 73 of the Constitution Act 1889 WA provides that the legislature of the State "shall have full power and authority, from time to time, by any Act, to repeal or alter any of the provisions of this Act". However, it is not lawful to present to the Governor for assent by the Queen [506] : "any Bill by which any change in the Constitution of the Legislative Council or of the Legislative Assembly shall be effected, unless the second and third readings of such Bill shall have been passed with the concurrence of an absolute majority of the whole number of the members for the time being of the Legislative Council and the Legislative Assembly respectively."
Section 73(2) was inserted into the Constitution by an amendment enacted in 1978. That subsection further restricts the capacity of the legislature to enact a Bill which expressly or impliedly provides for the abolition of either the Council or the Assembly (s 73(2)(b)) or provides that either House "shall be composed of members other than members chosen directly by the people" (s 73(2)(c)). The subsection requires that such a Bill should not be presented for assent unless "the second and third readings of the Bill shall have been passed with the concurrence of an absolute majority of the whole number of the members for the time being of the Legislative Council and the Legislative Assembly, respectively" (s 73(2)(f)) and the Bill has been approved by a majority of electors of the State at a referendum (s 73(2)(g); see also s 73(2)(e), s 73(3), (4), (5), (6)).
It is common ground that the Bill for the 1987 Act was presented for assent without there having been taken the steps described in s 73(2) of the 1889 Act. Hence, the submission by the plaintiffs on this branch of the case.
Before turning to the construction of s 73(2)(c), it is appropriate first to consider the authority of the legislature which enacted it to control the law-making processes of its successors.
It is necessary to begin with s 5 of the Colonial Laws Validity Act 1865 Imp (the 1865 Act). So far as relevant, this stated:
[Every] representative legislature shall, in respect to the colony under its jurisdiction, have, and be deemed at all times to have had, full power to make laws respecting the constitution, powers, and procedure of such legislature; provided that such laws shall have been passed in such manner and form as may from time to time be required by any Act of Parliament, letters patent, Order in Council, or colonial law for the time being in force in the said colony. The terms "representative legislature" and "colonial law" were defined respectively in s 1 as signifying "any colonial legislature which shall comprise a legislative body of which one half are elected by inhabitants of the colony" and as including "laws made for any colony either by such legislature as aforesaid or by Her Majesty in Council". It has been observed (by La Forest J) that the reference to "manner and form" appears in the proviso and that s 5 was aimed at validating not invalidating colonial laws [507] , and (by French J) that, while giving colonial legislatures power to alter their constitutions, s 5, paradoxically, also gave them a tool to bind their successors [508] . Observance of "manner and form" has been treated, notably in AG (NSW) v Trethowan [509] , as a condition upon which the "full power" referred to in s 5 is exercisable. 1. Mercure v The Queen [1988] 1 SCR 234 at 279(1988) 48 DLR(4th) 1 at 65. 2. The Hon Justice French, Manner and Form in Western Australia: An Historical Note, University of Western Australia Law Review, vol 23 (1993) 335, at p 338. 3. [1932] AC 526 at 539-540.
Section 2(1) of the Statute of Westminster 1931 Imp provided that the 1865 Act "shall not apply to any law made after the commencement of this Act by the Parliament of a Dominion", the latter expression including the Commonwealth of Australia but not the States. The Australia Acts, which came into operation on 3 March 1986, amended (by s 14) the 1889 Act, but not in any way presently relevant.
The Act of the Parliament of the United Kingdom known as the Colonial Laws Validity Act 1865 shall not apply to any law made after the commencement of this Act by the Parliament of a State.
The result is to remove any ground which might otherwise have existed for the submission that the 1987 Act was invalid for failure to meet any requirement of s 5 of the 1865 Act that its passage should have been in the manner and form required by the 1978 Act.
However, any gap left by the repeal of s 5 of the 1865 Act by s 3(1) of the Australia Acts was then filled by s 6 of the Australia Acts. This states:
Notwithstanding sections 2 and 3(2) above, a law made after the commencement of this Act by the Parliament of a State respecting the constitution, powers or procedure of the Parliament of the State shall be of no force or effect unless it is made in such manner and form as may from time to time be required by a law made by that Parliament, whether made before or after the commencement of this Act.
It will be noted that, in contrast to the form taken by the old s 5, s 6 is a direct statement of the mandatory requirement to observe manner and form. In its opening phrase, it is expressed to apply notwithstanding the declaration and enactment in s 2 of plenary legislative competence (save as to preservation of the status quo concerning relations with countries outside Australia) and in s 3(2) of power to repeal or amend Imperial law and to make laws repugnant to the law of England.
Further, it is submitted that the concluding words of s 106 of the Constitution, "until altered in accordance with the Constitution of the State", empower a State Parliament to insert into its Constitution a manner and form requirement, provided it is not inimical to the principles of representative democracy. This produces the result that the double entrenchment effected by s 73 is effective also by force of the statement in s 106 of the Constitution that the "Constitution of each State of the Commonwealth shall, subject to this Constitution, continue".
It may be accepted that the question whether a law of the State was enacted in a manner and form which did not satisfy s 73 would arise under the Australian Constitution, and in particular under s 106 thereof, for the reasons given by Burt CJ in Western Australia v Wilsmore [510] . This would be because the phrase in s 106 "altered in accordance with the Constitution of the State" recognises or accepts requirements, such as s 73, which otherwise apply to State constitutional changes. Accordingly, the sense of the phrase "altered in accordance with" would be "so altered as not to contravene any otherwise binding requirement". It would be a distinct question whether s 106 goes further so as to create any additional and binding category of restraint upon State constitutional alteration [511] .
There is a conceptual difficulty, to my mind, with the legitimacy of a manner and form requirement which is inserted in a written constitution otherwise than by a law made with observance of that manner and form which is thereafter to apply, or by a law having paramount force. Nevertheless, it is contended, primarily by the Solicitor-General of New South Wales, that there is a principle reflected in Bribery Commissioner v Ranasinghe [512] that a legislature has no power to ignore the conditions for law making imposed by the instrument which itself confers its power to make that law and that a law which changes those conditions must observe them.
Ranasinghe concerned the then Constitution of Ceylon which was conferred not by Statute of the Imperial Parliament but by The Ceylon (Constitution) Order in Council 1946. This preceded the Ceylon Independence Act 1947 Imp [513] . That statute denied the application of the 1865 Act to any law subsequently made by the Parliament of Ceylon but did not contain any equivalent of s 6 of the Australia Acts [514] . The case may stand for the propositions that a manner and form provision which appears in the written constitution of a unitary State where no paramount law, such as s 5 of the 1865 Act, remains in force, continues to place a restraint upon law making, and that the question of the observance of the restraint is justiciable; see Victoria v The Commonwealth and Connor [515] and West Lakes Ltd v South Australia [516] . It is unnecessary to resolve the matter in the present case. This is because, as I have indicated, whilst s 2(2) of the Australia Acts declares and enacts that the State Parliaments have plenary legislative power, it is further provided in s 6 that, notwithstanding this provision, manner and form requirements must be satisfied. This express treatment of the subject must leave no room for any greater operation which a principle derived from Ranasinghe might otherwise have had for any Parliament of an Australian State.
Accordingly, in my view, the presently binding effect of s 73 is derived from s 6 of the Australia Acts [517] . It was not, as I understand it, seriously contested first that the Bill for the 1987 Act was one respecting the constitution, powers or procedure of the Parliament of the State, within the meaning of s 6 of the Australia Acts and, secondly, that s 73(2), again within the meaning of s 6, placed a requirement as to manner and form of the making of laws of the description identified in it [518] . In any event, I am prepared to approach the case on the footing that both propositions are satisfied.
The central question then becomes whether, within the meaning of s 73(2)(c), the Bill for the 1987 Act was one which expressly or impliedly provided that either House should be composed of members other than members chosen directly by the people.
In Stephens [519] , Mason CJ, Toohey and Gaudron JJ spoke as follows of s 73(2):
[Section] 73(2) was plainly enacted with the object of reinforcing representative democracy and placing a further constitutional impediment in the way of any attempt to weaken representative democracy. And, so long, at least, as the Western Australian Constitution continues to provide for a representative democracy in which the members of the legislature are "directly chosen by the people", a freedom of communication must necessarily be implied in that Constitution, just as it is implied in the Commonwealth Constitution, in order to protect the efficacious working of representative democracy and government. 1. (1994) 182 CLR 211 at 233-234.
Against that background, there is a certain irony in the submissions now made by the plaintiffs. Plainly, the effect of the 1987 Act was to ameliorate the alleged malapportionment of which they complain. But the effect of their submission is that s 73(2) made it more difficult after 1978 for the legislature to achieve such a result, by placing and doubly entrenching an impediment in the path of the Bill for a statute such as the 1987 Act. The impediment was "doubly entrenched" because par (e) of s 73(2) adds to the list of Bills which require a particular majority in both Houses and submission to referendum a Bill which "expressly or impliedly in any way affects [section] 73".
Thus, the submissions for the plaintiffs would achieve a result opposite to that seen in Stephens as the object of s 73(2), namely, the object of placing a constitutional impediment in the way of any attempt to weaken representative democracy.
Further, the legislative intent in enacting s 73(2)(c) was not to affect the question of redistribution of electoral boundaries. This point was made on the second reading speech in the Legislative Council [520] . In the second reading speech on the Bill in the Legislative Assembly, the Premier had identified as one of the objectives of the legislation the forestalling of attempts "to by-pass the right of the electors at large to elect the members of either House" [521] .
It may readily be conceded, and the contrary is not contested, that the Western Australian Constitution provides for representative government. However, in the absence of any relevant entrenchment, effect is given to that doctrine by a fluid rather than a fixed constitution. In that respect there is a marked difference between the position in Western Australia and that which obtains under the Australian Constitution.
The thrust of the submissions for the State is that this fluidity permits a process of legislative change to the franchise and to the methods for composition of each chamber of the Parliament. There remains the Parliament of the State, the mere continued existence of which says nothing as to the nature of the franchise or the method of representation of the electors in the legislature.
I have referred earlier in these reasons to the taking of various steps over the last century by which the Parliament has pursued a path of change to the franchise and methods of representation. The point may further be illustrated as follows. At the elections for the Legislative Assembly held in 1897, the last before federation, the ratio (defined in the case stated as "the electoral ratio") between the largest number of electors in a constituency and the smallest number of electors in a constituency, divided by the number of members ordinarily returned by the constituencies was 38.52:1. The last election for the Legislative Assembly before the enactment of the 1978 Act took place in 1977. By then there were fifty-five constituencies, each returning one member, and the electoral ratio was 8.93:1. At the 1993 general election, conducted pursuant to the legislation under challenge in these proceedings, the electoral ratio was 2.91:1.
In Dixon v Attorney-General (British Columbia) [522] , the Court considered comparable issues by looking at deviations per constituency from a norm representing the average populations for each elected representative. I have earlier referred to the similar method adopted in the United States cases with deviations from "the ideal". The case stated does not give these figures upon an ideal (using either population at large or electors), but it may be supposed that, had it done so, a like trend over the last century would have appeared.
The last election for the Legislative Council held before federation took place in 1900. The electoral ratio was 22.22:1. At the 1977 election for the Legislative Council, the last before the introduction of the 1978 Act, the electoral ratio was 14.79:1, and at the 1993 general election the electoral ratio was 3.42:1.
Counsel for the State submits that the result of acceptance of the submissions for the plaintiffs would be to establish s 73(2)(c) as a barrier to further such development as regards apportionment of constituencies and electoral regions between the electors. This result, so the State contends, will be avoided and effect be given to the true interpretation of par (c) of s 73(2) if it be construed as "entrenching" only so much of the full elements of the system of representative government in the State of Western Australia as provide for popular election of members of each chamber in contrast to a process of nomination by the executive or of indirect election.
These submissions should be accepted. Section 73(2)(c) was not designed to restrict further change after 1978 to what the plaintiffs complain of as electoral malapportionment. The legislation assumed that the constitutional arrangements then in force in Western Australia were those for a representative government. It sought to protect this state of affairs (including the office of Governor, and the preservation of the bicameral legislature) in various respects as specified in s 73(2). But the legislature did not impose a manner and form requirement upon the changes to the electoral system such as those later brought about by the 1987 Act. The phrase "chosen directly by the people" was selected rather than "directly chosen by the people" as it appears in ss 7 and 24 of the Australian Constitution. This emphasises that the phrase is inseverable and conveys the one concept. That is, the entrenchment of the present system of popular selection of legislators by direct vote rather than by other indirect methods.
McGinty
Western Australia
(1926) 37 CLR 393
(1975) 135 CLR 1
(1992) 177 CLR 1
(1992) 177 CLR 106
(1994) 182 CLR 211
(1920) 28 CLR 129
(1947) 74 CLR 31
(1971) 122 CLR 353
(1945) 71 CLR 29
(1994) 182 CLR 104
(1979) 145 CLR 172
(1975) 135 CLR 337
(1979) 145 CLR 246
(1931) 44 CLR 394
(1960) 105 CLR 214
(1930) 44 CLR 319
(1983) 158 CLR 1
(1985) 159 CLR 192
(1995) 184 CLR 188
(1990) 169 CLR 482
(1994) 182 CLR 272
(1994) 181 CLR 96
(1904) 1 CLR 91
(1912) 15 CLR 18
(1943) 68 CLR 133
(1952) 85 CLR 159
(1962) 107 CLR 381
(1987) 162 CLR 574
(1982) 149 CLR 79
(1908) 6 CLR 309
(1964) 110 CLR 353
(1979) 143 CLR 190
(1981) 146 CLR 559
(1989) 168 CLR 461
(1993) 177 CLR 541
(1935) 54 CLR 262
(1965) 112 CLR 206
(1907) 5 CLR 201
(1975) 134 CLR 298
(1976) 135 CLR 552
(1989) 166 CLR 518
(1989) 99 FLR 265
(1995) 12 WAR 392
(1989) 85 ALR 621
(1915) 20 CLR 148
(1925) 37 CLR 252
(1964) 113 CLR 207
(1983) 153 CLR 297
(1936) 55 CLR 1
(1949) 79 CLR 43
(1904) 1 CLR 329
(1908) 6 CLR 469
(1949) 79 CLR 121
(1962) 108 CLR 372
(1968) 117 CLR 390
(1976) 135 CLR 110
(1913) 17 CLR 261
(1977) 139 CLR 585
(1994) 181 CLR 18
(1975) 134 CLR 201
(1977) 139 CLR 527
(1983) 152 CLR 254
(1961) 109 CLR 105
(1918) 26 CLR 9
(1924) 35 CLR 69
(1932) 47 CLR 482
(1956) 94 CLR 254
(1957) 95 CLR 529
(1965) 114 CLR 226
(1926) 38 CLR 380
(1992) 174 CLR 455
(1972) 128 CLR 221
(1986) 161 CLR 556
(1975) 134 CLR 81
(1980) 25 SASR 389
(1934) 51 CLR 518