this anomaly were a Beaconsfield administration, a Beaconsfield policy in the Eastern Question, an Afghan war, and an almost complete stagnation of domestic legislation for the six years from 1874 to 1880. With this startling result in view, the vital importance of modifying our electoral system in the direction of proportional representation scarcely needs further argument.

In 1880 the Liberals, with a majority of 464,000 votes, secured a majority of 178 members, when proportionately they were only entitled to 92. Thus the Conservative reaction of 1874 and the return to Liberal principles in 1880 were both exaggerated at these elections respectively, the Conservatives being over-represented at the former by 56 members, and the Liberals at the latter by 43. If Parliament then is at present in some sense a mirror of the body of electors, it is at any rate not a perfectly plane mirror, but one that largely magnifies and distorts the variations in political feeling in the country.

The circumstance of a majority of members in 1874 corresponding to a minority of votes is due to the fact that in majority representation majorities are counted and not weighed; so that, for instance, a majority in Birmingham of some 12,000 or 13,000 Liberals counted no more than that of 900 or 1,000 Conservatives in North Warwickshire, not to mention the tiny majorities of tens, twenties, or forties in such petty constituencies as Brecknock, Evesham, Marlow, Northallerton, Petersfield, or Thirsk. Other striking instances of the same kind occur in the election of 1880, where a majority of 2,740 for Sir Charles Dilke at Chelsea is balanced by a majority of 101 for Mr. Gorst at Chatham, and where a Liberal majority of just two at Colchester is balanced by a Conservative majority of 192 in East Essex.

It may be argued that by the abolition of the smaller constituencies or their absorption in larger ones many of these anomalies will be removed, and doubtless to some extent it will be so, for the smaller the constituencies the smaller on the average will be the majorities which turn the scale of the political balance. But even among enlarged constituencies there will be some where the predominance of one party is very large, and others where parties are pretty evenly balanced, and the large majorities of the former being counted as of no greater weight than the small majorities of the latter, large and uncertain deviations from proportional representation may be expected, if the system of majority representation remains unmodified. To cite but one instance in illustration: at the election of 1880 the Liberal majority of 1886 in South Durham exceeds the aggregate of the Conservative majorities, namely 1731, in all three of the divisions of Essex.

Descending from the aggregates for the whole United Kingdom, let us next examine the aggregates of a well-defined and fairly homogeneous group of constituencies, the counties of England and Wales, by the light of the following table:—

Upon this it is obvious to remark that, though we have here no instance of a majority of seats corresponding to a minority of votes, the Conservative majority of members elected has at each election been much larger than the respective strengths of the two parties justified; and that the Conservative reaction of 1874 enormously exaggerated this disproportion, giving the Conservatives an excess of forty-five seats, which, if they had been transferred to the rival party, to which of right they belonged, would have more than annihilated the Conservative majority of that Parliament.

Among the counties are included seven which are represented each by three members, and in these three-cornered' constituencies each elector can vote for two candidates only. The result is that they returned at the elections of 1868 and 1880 thirteen Conservatives and eight Liberals, and at that of 1874 fourteen Conservatives and seven Liberals, the Conservative reaction at the latter date being thus represented by the transfer of one seat only from the Liberal party to the Conservative. When we remember that these seven counties are Berkshire, Buckinghamshire, Cambridgeshire, Dorsetshire, Herefordshire, Hertfordshire, and Oxfordshire, all fairly typical agricultural counties without the disturbing influences of mining or manufacturing populations, it can hardly admit of a doubt that, had the rest of the counties been organised in like manner into three-cornered constituencies, the ratio of Conservatives returned to Liberals would have · been something like the following:-

116 Conservatives to 71 Liberals, instead of 135 Conservatives to 52 Liberals in 1868.

Thus, though the Conservatives would still have been unduly represented at all three elections in proportion to their actual strength, the deviation from an equitable proportion would have been much less than was actually the case, while in the election of 1874 the Conservative majority for the counties of England and Wales would have been reduced from 123 to 63—a reduction almost enough in itself to have annihilated the Conservative majority in the whole Parliament.

Such is the result of the system which Mr. Bright is never tired of holding up to ridicule and scorn as highly unconstitutional,' the bantling of Lord Cairns, supported by a small handful of sentimental VOL. XV.-No. 84.

X

:

and fancy statesmen and legislators sitting on our side of the House. Does it not afford the proof of its advantage to the country,' which Mr. Bright demands, after which only he will allow that we should permit any departure from the old lines of the Constitution, which, whatever its failings, we have a right to regard with some degree of reverence and affection'? Surely, if Mr. Bright had realised that adhering (in his sense) to the old lines of the Constitution meant a Beaconsfield administration for six years, while the proposed deviation from them meant merely a certain amount of difficulty and inconvenience in conducting the election at Birmingham so as to secure the return of three Liberal members for that city, he would have somewhat modified his essentially Tory 'attitude of reverence and affection for them. If everybody admits,' as Mr. Bright says, • that the present system (of voting in three-cornered constituencies) is an injustice and a failure,' it can only be because their attention has been more directed to the defects and awkwardness of the machinery than to the character of the results. The moral is, to amend and improve the machinery, not to fall back on a system which by its very nature must yield unfair and uncertain results.

Descending further from the aggregate of the county representation, let us select for examination four contiguous constituencies, each returning two members, the four divisions of the county of Lancashire. The results of the three elections since 1867 are summarised as follows:

Hence it appears that for twelve years from 1868 to 1880 the county of Lancashire had not a single Liberal representative, although 46 to 47 per cent. of the electors were Liberals, and that the smal increase to about 494 per cent. in 1880 sufficed to divide the representation equally between the two parties. Had the county been divided into two constituencies returning four members each, with the limitation that no elector could vote for more than three candidates (as in the City of London), the Liberals would certainly have had at least two representatives in 1868 and 1874, while in 1880 perhaps the most probable result would have been three Liberals to five Conservatives. This latter result would perhaps be more satisfactory to the sticklers for the rights of majorities than the present equal division, which, according to the singular mode of arguing they often adopt, leaves Lancashire entirely unrepresented, while upon their principles a system which gives the entire monopoly of

the representation to one party for twelve years is as just and successful as that adopted from Lord Cairns is an injustice and a failure.'

As our last group for examination let us take together the metropolitan constituencies in Middlesex, including the county of Middlesex, the City of London, and the six boroughs, Chelsea, Finsbury, Hackney, Marylebone, Tower Hamlets, and Westminster, an aggregate of about 272,000 electors in 1880. In the following table, besides the proportional and actual distribution of members, the distribution which would have resulted if the City of London had not elected its members subject to Lord Cairns's clause is shown:

Here the most striking feature is the over-representation of Liberals in 1868, converted by the return swing of the pendulum to a small over-representation of Conservatives in 1874, and again brought back to a small over-representation of Liberals in 1880. Thus the Conservative reaction of 1874, which under a system producing proportional representation would fairly have been measured by the transfer of two seats, actually resulted in the transfer of seven, and had not Lord Cairns's clause been in force in the City, would have resulted in a transfer of nine seats.

This invites our attention to a consequence of the existing system of majority representation, which is perhaps a still more serious evil than the anomalies hitherto shown to result from it, though it has been less frequently pointed out and less emphatically insisted on. This consequence is the instability of the representation caused by the shifting of small majorities in nearly balanced constituencies.

If the beam of a balance be supported at a point very near to its centre of gravity, the shifting of a small weight determines its inclination to this side or that. The system of majority voting has an analogous action: it balances those of the electors who have serious political convictions and hold them strongly-the steady Liberals against the staunch Conservatives; and then, if their weights are nearly equal, the inclination of the beam of the political balance is entirely at the mercy of a small body of electors, whose political views are determined at best by some ephemeral cry, some clever catchword, some panic fear, or some class interest, or in too many cases by those baser considerations which it may be hoped the Corrupt Practices Act of last Session will have done something to restrain.

This instability was largely exemplified in the election of 1874, when the current was temporarily in favour of the Conservatives, and perhaps to a nearly equal extent in that of 1880, when the set was in the opposite direction. At this latter election it was found that 37 seats were gained by the Liberals by a gross majority of 1,742 votes, an average of 47 each; and, not counting Irish seats, 47 constituencies changed sides with majorities of less than 100, six only of these being constituencies with less than 1,000 electors, where the deciding majority might perhaps be considered as a sensible fraction of the whole.

We may sum up the results of our examination of the working of the existing system of election since the Reform Act of 1867 in the following propositions :

1. Majority representation, merely counting majorities and not. weighing them, does not secure that a majority of electors shall always command a majority of representatives. (Witness the election of 1874.)

2. The results of majority representation generally deviate widely from the ideal-proportional representation.

3. In large groups of generally like constituencies, majority representation gives an excessive preponderance in the representation to that party which has the majority. (Witness the counties of England and Wales.)

4. Majority representation is unstable. Small shifting majorities have an undue influence on the representation, enormously exaggerating the fluctuations of political opinion in the country at large.

The Conservative reaction of 1874 was expressed out of all proportion to its real strength by the transfer of 81 seats from the Liberal to the Conservative party; and the opposite reaction of 1880, which brought back 120 seats to the Liberal party, doubtless received therein an exaggerated expression.

6

Though these defects in the system of majority representation do and must exist, it may be, and has in fact been, contended that they are practically neutralised by the particular conditions under which it is actually exercised. The facts and figures we have cited show that this is not the case, though it may be admitted that the variety in the franchise, and the variety in the constituencies which are included within the old lines of the Constitution,' have had a considerable, though uncertain, mitigating effect. Much was done in 1832 and in 1867 to bring the franchise nearer to uniformity. The rights of suffrage of the scot-and-lot voters, the potwallopers, the freemen, the common-councilmen, the non-resident freeholders in boroughs, &c., have been taken away or merged in some wider qualification. The distinction between the qualifications in counties and boroughs survives, but it is doomed, and the coming Session will have