Proportional Representation: A Study in Methods of Election

by John H. Humphreys

APPENDIX I

THE JAPANESE ELECTORAL SYSTEM—THE SINGLE NON-TRANSFERABLE VOTE

The following memorandum has been written by Mr. Kametaro Hayasbida, the Chief Secretary of the Japanese House of Representatives, in reply to a series of questions, the particulars of which are set out in the memorandum.

Failure of single member system.

The Original Election Law of our country was promulgated in 1889, the same year in which took place the promulgation of the Constitution. Under this law the system of small electoral districts was single-adopted, and each Fu or Ken (administrative district) was divided into several electoral districts each of which constituted a single-member constituency (with the exception of some large districts which, impossible of further division, had two seats allotted with the system of scrutin de liste). The system was, however, found in practice to be very unsatisfactory, as it often happened that a minority of the voters, instead of the majority, in certain Fu or Ken obtained the majority of the members returned, and, on the other hand, a party with a majority at the polls could not sometimes, as the result of the grouping of the voters in the small electoral districts, secure any representation at all. Under such circumstances it was utterly impossible for each political party to obtain representation in reasonable proportion to the strength of its voters; or, in other words, the electors of the country at large had never succeeded in being properly represented in their legislative body. As the inadequacy of the system was thus apparently shown I formulated in 1891, by somewhat what modifying Marshal's cumulative voting system, a system of large electoral districts combined with that of the single vote, and urged for a revision of the Election Law.

Multi-member constituencies. Single vote adopted 1900.

Since then several elections had taken place; and the defects of the existing law were more strongly pronounced at each successive election. It was, however, not until the year 1898 that the Government at last introduced a Bill for a revision of the law with the view of adopting the system I had the honour of formulating. After heated discussion in three successive sessions, the Bill was passed in 1900 and sanctioned as a law. This is our present Election Law. In the revised system the Fu, Ken, and Shi (the administrative districts) constitute at the same time the electoral districts, and a voter in each district has but one vote for one candidate, while several seats (according to the population) are allotted to the district.

The above is a brief historical sketch of our electoral system. I shall now try to answer your questions in order.

Equitable results.

As to the first question whether our system secures the representation of each party in reasonable proportion to its voting strength, I cannot do better than answer it by pointing out a few instances in the General Election which took place on the 15 May 1908.

TABLE I

THE CITY OF TOKYO (11 seats)

Seats in Seats
Parties. Votes. Proportion Obtained.
to votes.
Seiyu-Kwai (Liberals) 6,579 2.71 2
Konsei-honto (Progressives) 2,216 0.91 1
Daido-ha (Conservatives) 2,879 1.18 2
Yuko-Kwai (Radicals) 4,656 1.91 2
Churitsu (Independent) 10,414 4.29 4
——— ——- —
Total 26,744 11.00 11

All parties except the Seiyu-kwai and Daido-ha succeeded in obtaining their representatives in reasonable proportion to their respective voting strength. The explanation given for the particular case of the Seiyu-kwai is that the party, unable for some reason or other to limit the number of candidates, had placed five candidates instead of three or four, and caused its own defeat by splitting the votes. I take at random, or rather in the order they come, a few more districts, and the results obtained are as follows:—

TABLE II

TOKYO-FU (5 seats)

Parties. Number of Seats in Seats
Candidates. Votes. Proportion Obtained
to Votes.
Seiyu-kwai 5 12,794 4.02 4
Kensei-honto - - - -
Daido-ha. 1 13,122 .98 1
Churitsu - - - -
——— —— -
Total 6 15,916 5.00 5

TABLE III

THE CITY OF KYOTO (3 seats)

Parties. Number of Seats in Seats
Candidates. Votes. Proportion Obtained
to Votes.
Seiyu-kwai 1 1,284 0.45 -
Kensei-honto - - - -
Daido-ha - - - -
Yuko-Kwai - - - -
Churitsu 3 7,304 2.55 3
- ——- —— -
Total 4 8,588 3.00 3

TABLE IV

KYOTO-FU (5 seats)

Parties. Number of Seats in Seats
Candidates. Votes. Proportion Obtained.
to Votes.
Seiyu-kwai 5 18,928 4.01 4
Kensei-honto — — — —
Daido-ha — — — —
Yuko-kwai — — — —
Churitsu 1 4,701 0.99 1
———————————————————
Total…. 6 23,629 5.00 5

TABLE V

THE CITY OF OSAKA (6 seats)

Parties. Number of Seats in Seats
Candidates. Votes. Proportion Obtained.
to Votes.
Seiyu-kwai 5 8,666 3.32 4
Kensei-honto — — — —
Daido-ha — — — —
Yuko-kwai 1 2,612 1.00 1
Churitsu 2 4,368 1.68 1
——————————————————————-
Total…. 8 15,646 6.00 6

TABLE VI

OSAKU-FU (6 seats)

Parties. Number of Seats in Seats
Candidates. Votes. Proportion Obtained.
to Votes.
Seiyu-kwai 5 15,137 3.57 5
Kensei-honto — — — —
Daido-ha 1 2,199 0.52 —
Yuko-kwai 1 1,304 0.31 —
Churitsu 3 6,786 1.60 1
——————————————————————-
Total…. 10 25,426 6.00 6

Throughout all electoral districts similar results were obtained. The Churitsu (i.e. those belonging to no party), considered as a group, had not everywhere been as successful as the other parties, as observe in Tables V. and VI. Each candidate of this group is quite independent of the other, and has no political views or propaganda in common, nor any organization whatever. Therefore, each case is totally different from the other. Although all independent candidates or voters are in these tables grouped as Churitsu, it is not proper to consider them in the same category with the other parties.

Now, judging from the results in the General Election, a few instances of which are given above, I may say that our present system, if not fully satisfactory, tolerably secures the representation of each political party in approximate proportion to its voting capacity.

The new system and party organization.

As to the first part of your second question, whether, to obtain these results, the system involves a great deal of calculation on the part of political organizations as to the exact number of their supporters, I should say that, as the same system and method of election are uniformly adopted in the city, county, borough and village elections as well as in the elections of the Prefectural Assembly, it is not a very difficult task for all political parties to ascertain from the results of all these elections their relative strength, and to estimate the number of their supporters.

As to the second part of the question, whether it is necessary to issue precise instructions to the electors as to the candidates for whom they should vote, my answer is this: as every political organization through its branch in every Fu and Ken and the sub-branches in the cities, counties, towns and villages, is always in close touch with its constituents, and is constantly explaining its position and propaganda, with the view not only to instruct them but also to extend the sphere of its influence, it is not so difficult as it seems to decide the number of candidates. When it is once decided efforts are made on the part of the organization to distribute the votes among the candidates in such a way that not one of them receives a defeat at the hands of the other party. To attain this object the methods are not very complicated, for every elector has but one vote for one candidate; and, moreover, the stronger candidates, so long as their own position is secured, will endeavour to distribute a portion of their votes among the weaker candidates. This being the case, the member returned with the greatest number of votes may not be the most popular candidate, but the party as a whole is much more likely to succeed in getting representatives in proportion to the strength of its voters.

The position of independents.

As to the third question, whether the system enables men of independent mind and character to maintain their position in Parliament, I should emphatically state that the revised system is much better than the old in this respect. Under the old system even such a prominent man as Mr. M. Matsuda (the Speaker of the House of Representatives some years ago, and the Minister of Finance in the present Government) suffered several defeats. But under the new system it has never happened that the leader of a party has lost his seat at any election, as he may seek his election at the safest district. To men of independent mind and character the new system offers the greater opportunity to maintain their position in the House, for in the election they may, in spite of the opposition of parties, draw their votes from all parts within a large electoral district. It may be said that the larger electoral district we have, the greater opportunity we afford to independent candidates. For instance, both Mr. Y. Ozaki, the Mayor of Tokyo, and Mr. S. Shimada, by being independent candidates, have never lost their seat in Parliament, and in the last General Election were returned for their native prefecture or town with a great number of votes.

This brings me to the end of my answers to your inquiries. In conclusion
I may say a few words about the public opinions in our country as to the
Election Laws.

Public opinion and the new system.

Despite the fact that the new system enables the elector of the country to be more reasonably represented in the House, still there are some ambitious politicians urging for their own selfish purpose to restore the old system. But, as almost all prominent members in both Houses are fully cognizant of the relative merits and demerits of the two systems, there is not much chance of our returning to the old system.

APPENDIX II

THE SECOND BALLOT

A Note on the German General Elections of 1903 and 1907.

The German Reichstag, which consists of 397 members, is elected by a system of single-member constituencies. Every member, however, must have obtained a majority of the votes polled, either at a first or second ballot, in the constituency for which he has been returned. The German Official Returns furnish very complete details of the elections, including the figures for the first and second ballots, and the summaries at the end of the Returns disclose a very striking divergence between the proportions of seats obtained and votes polled by the various political parties. These discrepancies have attracted general attention, and have usually been attributed to the great variation in the size of German constituencies. As a matter of fact, the effect of redistribution on the proportionality between seats and votes is not nearly so large as is generally supposed. Apart from the consequences of neglecting the votes of the minority or minorities in each constituency, wherein lies the gravest defect of a single-member system, the second ballot is a disturbing factor of considerable importance. So far from diminishing the disproportion between seats and votes polled by the various parties, the second ballot frequently increases that disproportion. In order to appreciate the respective effects of unequal constituencies and of the second ballots it is necessary to consider these two factors separately. This will be facilitated by making a comparison between the results which would have been obtained without second ballots with the results actually obtained. The following tables, which are based upon the official returns, give the votes polled and the seats obtained by the five principal groups:—

GERMAN GENERAL ELECTION, 1903

Parties. Votes. Results without Results with
Second Ballot. Second Ballot.
Social Democrats 3,010,771 122 81
(31.7%) (30.7%) (20.4%)
Centre Party 1,875,273 104 100
(19.7%) (26.2%) (25.2%)
National Liberals 1,317,401 32 51
(13.9%) ( 8.1%) (12.8%)
Conservatives 1,281,852 79 75
(13.6%) (19.9%) (18.9%)
Radical Parties 872,653 11 36
( 9.2%) ( 2.8%) ( 9.1%)

GERMAN GENERAL ELECTION, 1907

Parties. Votes. Results without Results with
Second Ballot. Second Ballot.
Social Democrats 3,259,029 73 43
(28.9%) (18.4%) (10.8%)
Centre Party 2,179,743 101 105
(19.3%) (26.4%) (26.4%)
National Liberals 1,630,681 47 54
(14.5%) (11.8%) (13.6%)
Conservatives 1,632,072 91 84
(13.6%) (22.9%) (21.2%)
Radical Parties 1,233,933 30 49
(10.9%) ( 7.6%) (12.3%)

The effect of unequal constituencies on representation.

The Social Democrats were affected to a greater extent than any other party by both the factors referred to. In 1903 the Socialists polled 31.7 per cent, of the votes, and, at the first ballots, were at the head of the poll in 122, or 30.7 per cent, of the constituencies. In other words, if the system of second ballots had not been in force, the Social Democrats would have obtained very nearly their fair share of representation. If, in addition, there had been a redistribution of seats by which the sizes of constituencies had been equalized, the Social Democrats would have obtained more than their share of representation. The strength of the party lay in the large towns, and if, for example, Berlin had the additional eight seats to which it was entitled nearly all of them would have fallen to the Social Democrats. Again the three divisions of the district of Hamburg returned Social Democrats with overwhelming majorities. Were the representation allotted to Hamburg doubled, as it should be, all six seats might possibly have fallen to the Social Democrats.[1] An equalization of the size of constituencies might have produced in 1903 the phenomenon which has occurred so often in England. The largest party would have secured a number of seats far in excess of that to which it was entitled by reason of its strength. In 1907 the Socialists polled 28.9 of the votes, but only succeeded in reaching the head of the poll at the first ballot in 73, or 18.4 per cent. of the constituencies. A redistribution of seats would have added to their representation in the large towns, and the first ballots would have yielded a result which would have corresponded more fairly with their polling strength.

The effect of second ballots.

In both years the system of second ballots has had the effect of reducing very considerably the representation of the Social Democrats. In the year 1903 the Social Democrats won 56 constituencies by absolute majorities, and were engaged in the second ballots in 118 constituencies. In 66 of these constituencies they were at the head of the poll, but succeeded in maintaining this position at the second ballots in 24 only. In the remaining 52 constituencies they were second on the poll, and at the second ballots they were able to win only one of these seats. In these 118 constituencies the Socialists polled 1,170,000 votes at the first ballots, whilst the other parties polled 1,920,000. As a result of the second ballots the Socialists obtained 25 seats and the remaining parties obtained 93 seats.

The figures of the year 1907 tell a similar tale. At the first ballots the Social Democrats were at the head of the poll in 73 constituencies. The second ballots reduced this number to 43. They were engaged in the second ballots in 90 constituencies; they were at the head of the poll in the first ballot in 44 of these constituencies, but kept this position in 11 only; they were second on the poll in the remaining 46 constituencies and won in 3 cases only. In these 90 constituencies the Social Democrats polled at the first ballot 1,185,000 votes, whilst the other parties taken together polled 1,888,000 votes; the Socialists obtained 14 seats, the other parties obtained 76 seats.

In both these elections the second ballots affected very adversely the representation of the largest party. If this party, without the second ballot and with a fair distribution of seats, might have obtained more than its share of representation, then the second ballots would have acted as a corrective, but not necessarily so. There is no reason why the second ballots should not have added to the over-representation already obtained. This will be seen from the figures of the elections in the Kingdom of Saxony. This division of the German Empire is entitled to 23 representatives in the Reichstag. In 1903 the Socialists won 18 of these seats with absolute majorities; they were engaged in the second ballots in the remaining five constituencies; they won four (all those in which they were at the head of the poll at the first ballots) and only lost the one constituency in which they were second on the poll. The Social Democrats, who at the first ballots polled 58.8 per cent, of the votes, thus obtained 22 seats out of 23, and the second ballots in this case only confirmed the overwhelming preponderance which the system of single-member constituencies had conferred upon the larger party.

Second ballots and the swing of the pendulum.] It would, indeed, seem that a system of second ballots rather accentuates those great changes in representation which are the normal characteristic of a system of single-member constituencies. In the elections in Saxony in 1907 the Social Democrats were still by far the largest party, obtaining 48.5 per cent. of the votes. They succeeded in obtaining eight seats by absolute majorities and were engaged at the second ballots in eight other constituencies. They lost every one of these constituencies, although at the first ballots they had been at the head of the poll in five of them. The unfavourable swing of the pendulum reduced their representation at the first ballots, and the second ballots merely increased their misfortunes.

Nor would redistribution have lessened the violence of these changes in the constituencies in which second ballots were necessary. Thus, for example, Frankfort-On-Main, with an electorate of 77,164, should return two members instead of one. The constituency was won by the Socialists in the second ballots of 1903, but was lost at the second ballots in 1907. In both years the Socialist candidate was at the head of the poll at the first ballots. Similarly the constituency of Elberfeld-Barmen, with an electorate of 67,241, won by an absolute majority in 1903, was lost by the Socialists at the second ballots in 1907, although their candidate had been at the head of the poll at the first ballot. If these and other constituencies had received additional representatives, the violence of the changes in the composition of the legislative body would in all probability have been increased.

The second ballot and the representation of minorities.

A study of the statistics of the German General Elections shows that the representation obtained by the various parties depends very largely upon their supremacy in certain localities. In these districts the minorities have been unrepresented for many years, the second ballots having in no way saved them from practical disfranchisement. Thus the Centre Party is in the ascendant in the Rhenish Provinces. In the district of Cologne, Münster, and Aix-la-Chapelle, the Centre Party monopolizes the representation, returning in 1907 every one of the 15 members to which the districts were entitled. In the adjoining districts of Dusseldorf, Coblentz and Treves they returned 16 out of 24. In Bavaria, the districts of Lower Bavaria, the Upper Palatinate, Lower Franconia and Schwabia, which are entitled to 23 members, were represented wholly by members of the Centre Party. Taking the kingdom of Bavaria as a whole, the Centre Party obtained 34 seats out of 48, although they polled only 44.7 per cent of the votes at the first ballots. There is therefore reproduced in Germany the conditions which obtain in certain parts of the United Kingdom—the permanent supremacy of one party which monopolizes, or nearly so, the representation of the district.

Summary

The system of second ballots has therefore had a considerable influence in creating that divergence between the votes polled and the seats obtained which has characterized German elections. The representation of any one party depends, to a very large degree, upon the attitude taken towards it by other parties. The system in no way acts as a corrective to the anomalies arising from single-member constituencies, and may even accentuate the violent changes associated with them. Moreover, the system does not provide representation for minorities, and therefore does not ensure a fully representative character to popularly elected legislative bodies. It may be mentioned that all the criticisms here directed against the second ballot apply with nearly equal force to the use of the alternative vote (see p. 95), a thinly disguised form of the same principle which appears to be meeting with some acceptance in this country.

[Footnote 1: The minority would, of course, have had a better chance with six divisions. Dr. Ed. Bernstein, to whom the author submitted this memorandum, makes the following comment: "I am not so sure that the equalization of the size of the constituencies would in 1903 have secured to the Social Democratic party a number of seats far in excess of its voting strength. But this is a subordinate consideration. The possibility of an unproportional representation of parties, even if the seats are equally distributed, is undeniably there, and this ought to settle the question.]

APPENDIX III

THE SWEDISH SYSTEM OF PROPORTIONAL REPRESENTATION

The principle of proportional representation was first discussed in Sweden in 1867. The new Danish Constitution of that year provided for the use of the transferable vote (Andrae's scheme) in the election of the Upper House, and Herr S. G. Troil proposed in the Swedish Parliament that the three most important of its committees should be elected by means of the same system. The motion was not carried, and a similar motion, made by Professor H. L. Ryön in 1878, was equally unsuccessful. It was not until 1896 that the next step was taken, when the Government, in view of the increasing demand for a more democratic franchise, proposed a proportional system of election. Nothing came of this proposal immediately, but from this date the agitation for an extension of the franchise gave rise to the demand for the proportional method of election in order to ensure the representation of minorities.

The former constitution of the two chambers.]

The story of the struggle for reform will best be understood if prefaced by a statement of the franchise conditions previously existing in Sweden. The Upper, or First, Chamber of the Riksdag, was elected by members of the provincial councils and of the councils of the five largest towns. The other towns sent members to their provincial councils. The members of provincial councils were elected in two stages; the primary electors chose electors of the second degree, who in turn chose the councillors. The primary electors in the country[1] had ten votes for every 100 kroner of rateable income, subject to a limit of 5000 votes. The electors of the second degree had only one vote in the election of councillors, and councillors had only one vote in the election of members of the First Chamber of the Riksdag. Owing to the great advantage conferred upon primary electors possessed of large incomes these electors largely controlled not only the composition of the town and provincial councils, but also the composition of the Upper Chamber. The election of members of the Lower Chamber of Parliament was direct; every person of not less than 800 kroner income was entitled to vote, but no one was entitled to more than one vote.

The struggle for electoral reform.

In 1899 M. Branting, the leader of the Socialist Party, proposed the adoption of proportional representation, coupled with universal and equal suffrage for the election of town councils. The main object of this proposal was to place town councils on a more democratic basis, but as the five largest councils elected representatives to the First Chamber the proposal would have had some influence upon the composition of that House. M. Branting's proposal was rejected, and when revived two years later met a similar fate. In 1902 two Liberals (MM. Hedlund and Carlsson) proposed that provincial councils should be elected by a proportional method on the basis of manhood suffrage, whilst a similar proposition was made in the same year in respect of the elections of the Lower House of Parliament. Both these motions were rejected, but in response to a demand from both Houses for an inquiry a Royal Commission was appointed to consider the problem of electoral reform. The Commission reported in the following year in favour of a list system of proportional representation with official ballot papers, and the Government proposed this system combined with manhood suffrage for the election of members for the Lower Chamber. This proposal was accepted in 1904 in the Upper Chamber, but rejected in the Lower Chamber by five votes. Next year it was again discussed, accepted by the Upper Chamber but rejected in the Lower by a majority of ten. A change of ministry took place, and in 1906 M. Staaff, the Liberal Prime Minister, proposed manhood suffrage with the "majority" system of election. But the Moderate Party insisted upon a proportional system, and the proposals of the Liberal ministry were rejected by the Upper Chamber. M. Alfred Petersson, of Paboda, then proposed manhood suffrage with a proportional system for the Lower Chamber, and a proportional system for the Upper Chamber, which, however, was to be elected as before by the provincial councils. This proposal was rejected by the Lower Chamber but accepted by the Upper Chamber, and M, Staaff resigned. The Moderates, with M. Lindman as Prime Minister, then introduced a Bill incorporating M. Petersson's proposals with the addition of the direct election of provincial councils and a less plutocratic franchise. This measure, which was adopted by both Houses in 1907, was confirmed after a General Election in 1909.

The Swedish law of 1909.

Under this law the proportional system is applied to elections for both Houses of Parliament, all parliamentary committees, town councils and provincial councils. For the Lower Chamber there is manhood suffrage. The Upper Chamber is elected still by the provincial councils and by the town councils of the five largest towns, but the elections of provincial councils are now direct. But, in order to maintain as much continuity as possible in the composition of the Upper Chamber, only one-sixth of the House is renewed every year. The maximum number of votes in the elections of both provincial and town councils is forty. The first election under the new system took place in 1909, when the Stockholm Town Council and several provincial councils were called upon to elect their proportion of members of the Upper House. In March 1910 the first elections to the Stockholm Town Council were held, and in the following May there were elections under the new system for all the provincial councils. In 1911 the first elections to the Lower House of Parliament will take place.

In Sweden, even under the new law, there are no official ballot papers and no nominations of candidates. This arrangement is supposed to preserve to the electors the fullest possible liberty in voting. In practice the party organizations print ballot papers containing the names of the candidates whom they support, and these printed forms are accepted by the returning officers. Every elector, however, is at liberty to strike out any of the names on these papers, to substitute other names, to vary the order in which the names are printed, or to prepare his own ballot paper.[2]

The Swedish system of proportional representation.]

The mechanism of the proportional system adopted has had regard to the practice mentioned in the preceding paragraph. The first proposal, that of M. Petersson, of Paboda, was only a crude approximation towards a proportional system. His scheme, in brief, was (1) that the number of votes recorded for each candidate should be ascertained; (2) that the candidate with the highest number of votes should be declared elected; (3) that a further count should then take place, the papers on which the successful candidate's name appeared being treated as of the value of one-half. The remaining candidates whose names appeared on these papers would be credited with half a vote in respect of each such paper. The non-elected candidates would then be arranged according to the number of votes obtained, the highest being declared elected. As soon as any two names on any ballot paper had been declared successful a fresh count would take place, such papers being treated as of the value of one-third. This process of reducing the value of the paper as soon as a further candidate appearing thereon was elected was to be continued until all the seats were allotted. The principle underlying this distribution of seats is the same as that contained in the d'Hondt rule of the Belgian system. A group of electors which was more than twice as numerous as any other group would obtain two seats before any was allotted to a smaller group. If the group was more than three times as large as any other it would obtain three seats before the smaller group received one, and so on. It was at once recognized that this scheme would tell considerably in favour of well-organized parties—parties whose supporters would accept the ballot papers printed for them without question. An example will make this clear. If, taking an extreme case, in an election for three members 8000 voters placed the names of two candidates, P and Q, on each of their ballot papers, whilst a more loosely organized group of 13,000 voters spread its support over four candidates, T, S, V and W, different sections voting for these candidates independently, the following result might take place:—

P Q . . 8,000 | T . . . 4,000 | S . . . 3,500 | V . . . 3,000 | W . . . 2,500

Candidate P, being the first in order on the 8000 ballot papers of the first group, would be declared elected, and Q, the remaining name on these ballot papers, would be credited with 4000 votes—half the original value of the papers. Q and T, having 4000 votes each, would then be declared elected. Thus one group, with 8000 votes, would carry two seats, and the other, with 13,000 votes, would only obtain one—a result due to a lack of combination.

The allotment of seats to parties.

The plan finally adopted is based on M. Petersson's proposal, but provides, as in the Belgian scheme, for the official recognition of parties. Electors may write at the head of their ballot papers the name or motto of a party. The papers bearing the same name or motto are then grouped together, the numbers in each group ascertained, and the seats available are allotted to these groups in accordance with the d'Hondt rule, irrespective of the number of votes obtained by individual candidates. Thus, in the example given, if electors of the second group had all headed their ballot papers with the same party name or motto the particular way in which they had distributed their votes among the candidates would not have affected the number of seats obtained by the group as a whole. The first group would have obtained one, and the second two seats.

The selection of the successful candidates.

The position of the candidates on each list is determined in accordance with the original proposal of M. Petersson. The candidate receiving the highest number of votes is declared elected, the papers on which his name appears are then marked down to the value of one-half, the relative position of the remaining candidates ascertained afresh, and the highest of these declared elected, and so on. This procedure, called the reduction rule, is however subordinate to a further rule (the rule of the order of preference), which is as follows. If more than one-half of the supporters of a party list have placed the same candidate at the head of their ballot papers, the first seat apportioned to the list is allotted to this candidate; if more than two-thirds have placed the same two candidates in the same order at the head of the ballot papers, these two candidates have the first claim to the seats apportioned to the party; if more than three-fourths have placed the same three candidates in the same order at the head of the list, these are given the first, second, and third seats, and so on. The selection of the successful candidates is determined in accordance with this rule so far as possible, but as soon as the application of the rule breaks down the relative claims of the non-elected candidates on the list are determined in accordance with the reduction rule. But if, say, three candidates have been declared elected in accordance with the rule of the order of preference, and it is necessary to choose others by the reduction rule, the papers containing these three names are treated as of the value of one-fourth in determining the relative position of the remaining candidates of the group.

Free voters and double candidatures.

In order to complete the description of the Swedish system two subsidiary features, which will seldom come into play in actual elections, must be mentioned. Provision is made for those electors who owe no party allegiance, and who therefore do not wish to place any party name or motto at the head of their list. Such voters are called "free voters," and the votes recorded for their candidates are ascertained. These candidates are placed in a group by themselves, called the free group, but the number of votes recorded for each individual candidate in this group, and not the total number of votes recorded for all the candidates, forms the basis of comparison with the totals of the party lists in the allotment of seats. The second feature provides for the improbable case of two groups of electors or parties having placed the same candidate upon their list. In the event of such candidate being so favourably placed in two lists as to be elected by both parties, then, for the purpose of ascertaining the new value of the papers on which his name appears, each list is debited with half a seat. When, as already explained, one seat has been allotted to a list, the list total is divided by two in accordance with the d'Hondt rule for the purpose of the fresh comparison of totals; but if this candidate has already been elected on another list the total would be divided by one and a half instead of by two. A fresh total would be ascertained for each of the lists containing the candidate's name.

An election at Carlskrona.

The author was permitted by the courtesy of the Burgomaster of Carlskrona to watch the election of provincial councillors on 24 May 1910, to represent the city in the Bleking provincial council, and a description of this election will show how the system works in practice. Carlskrona is entitled to nine members. For the purpose of the election the town was divided into two parts, but the polling place in each division was at the town hall. The register was prepared fourteen days before the election, and stated in addition to the name, address, and occupation of the elector, the amount of his (or her) rateable income and the number of votes to which he (or she) was entitled. The conduct of the election was in the hands of the Burgomaster, assisted by the magistrates of the town. As already explained, there were no official ballot papers and no nominations of candidates. Each elector voted for such candidates as he pleased, provided they possessed the necessary qualifications—those of an ordinary elector. Three parties—the Moderate, Liberal, and Labour—contested the election. Each party printed ballot papers containing the names of the candidates adopted by the party organization and with the name of the party at the head of the ballot paper. The ballot paper issued by the Moderate party was in the following form:—

De Moderata

Borgmästaren—O. Holmdahl. Grosshandlaren—N. P. Nordström. Lasarettsläkaren—R. Lundmark. Disponenten—H. Berggren. Kommendören—G. Lagercrantz. Rådmannen—C. G. Ewerlof. Chefsintendenten—I. Neuendorff. Kaptenen, friherre—F. E. von Otter. Underofficeren af 2: dra graden—O. W. Strömberg. Folkskolläraren—H. E. Mattsson. Byggmästaren—K. J. A. Johansson. Handlanden—Aug. Andrén.

The Poll.

The ballot papers could be obtained at the committee rooms on, or prior to, the day of election, and also on the day of election from party agents at the doors of the polling stations. Each elector took his ballot paper folded to the Burgomaster, or presiding magistrate, who endorsed the back with the number of votes to which the elector was entitled. The presiding magistrate was assisted by two others who checked the accuracy of the proceedings. The poll opened at 10 A.M., the proceedings were adjourned for lunch at 1 P.M., the poll was again opened during the afternoon and closed about 8 P.M. The counting took place next day when, as comparatively few electors took advantage of their right to vary the order of the names as printed on the ballot papers, the number of votes recorded for each candidate was easily ascertained. Nor did the varying values of the ballot papers present any great difficulty. A calculating machine made the necessary additions both quickly and accurately. In this election only one paper was spoiled,[3] and it was very obvious that the provision of printed ballot papers by the party organizations made the act of voting a very simple one. The votes recorded for the different parties were as follows:—

Moderate . . . . . 20,334
Liberal . . . . . 8,732
Labour . . . . . 3,617

_The allotment of seats to parties.

There were nine seats to be distributed among the three parties. The distribution was carried out in accordance the d'Hondt rule, but the method of applying this rule differed from that employed in Belgium. In Belgium the party totals would have been divided by the numerals 1, 2, 3, &c., and the quotients ranged in order of magnitude, the ninth in order being termed the "electoral quotient." Each party would have received as many seats as its total contained this quotient. The Swedish method provides for the allotment of one seat at a time, and it does so because of the possibility of the same candidate being elected by more than one party. Save in the rare case mentioned, the arithmetical operations, though differently presented, are identical with those of the Belgian system. Thus, at Carlskrona the first seat was given to the Moderates—that party having received the highest number of votes. Before the next seat was allotted the value of the Moderate total was reduced by one-half, and the new total was then compared with the original totals of the other parties. The totals to be considered in the allotment of the second seat were, therefore, as follows:—

Moderate. . . . . 10,167
Liberal . . . . . 8,732
Labour . . . . . 3,617

The Moderate party being still credited with the highest total received the second seat, and their original total, 20,334, was then divided by three in order to ascertain to whom the third seat should be allotted. The totals at this stage were as follows:—

Moderate . . . . . 6,778
Liberal . . . . . 8,732
Labour . . . . . 3,617

The Liberal total being now the highest, this party received the third seat, and in order to ascertain to whom the fourth seat should be given the Liberal total was reduced in value by one-half, the totals of the other parties remaining as at the previous allotment. The totals for comparison were now:—

Moderate . . . . . 6,778
Liberal . . . . . 4,366
Labour . . . . . 3,617

The Moderate total was again the highest, and the party received the fourth seat. The process of reducing the totals in succession according to the foregoing rule was continued until all the nine seats were allotted. In this election the Moderates obtained six seats, the Liberals two, and Labour one.

The selection of the successful candidates.

The returning officer had then to determine which candidates on each list should be declared successful. In the Carlskrona election this task was extremely simple, for the large majority of the voters had accepted the ballot papers provided for them by their parties. No less than 19,756 votes out of a total of 20,334 had been received for the Moderate list as printed by the party organization. The totals for each candidate were quickly ascertained. Moreover, it was possible to select all the successful candidates by the rule of the order of preference. More than six-sevenths of the Moderate votes having been recorded for the list as printed, the first six names on the list were declared elected. Of the Liberal votes, 8118 out of a total of 8732 were recorded for the party list as printed, and as this number constituted more than two-thirds of the total, the first two names on the list were declared elected. With regard to the Labour party, 3580 out of a total of 3617 votes had been recorded for the party list, and the first candidate on the list was therefore declared elected.

_The election of suppléants.

In common with all continental systems, supplementary members (suppléants) were chosen for the purpose of taking the place of an elected member who might die or retire before the council had run its course. The method adopted in Sweden is peculiar to itself. In Belgium the same rules serve for the election of the suppléants as for the election of members, and they are called upon to serve in the order in which they stand at the declaration of the poll. In Sweden it is held that each elected member must have a suppléant, or deputy, special to himself. The method of selection may be illustrated from the Carlskrona election. The candidate who was to be regarded as suppléant to Burgomaster Holmdahl (the first on the Moderate list) was chosen as follows: Holmdahl had received 20,334 votes, his name having appeared on every ballot paper of the Moderate party; the votes recorded for the unelected candidates on these papers were ascertained, the result being:—

Neuendorfs . . . . . 20,334
von Otter . . . . . 20,242
Strömberg . . . . . 19,913
Mattsson . . . . . 20,119
Johansson . . . . . 20,237
Andrén . . . . . . 20,170

Neuendorff being the candidate who had received the highest number of votes on these papers, was declared elected as suppléant to Holmdahl. A suppléant for Nordström, the second elected member, was then chosen from among the remaining five non-elected members. Nordström's votes were 20,235, and the votes recorded for the non-elected members on the same papers were:—

von Otter 20,143
Strömberg 19,913
Mattsson 20,055
Johansson 20,195
Andrén 20,071

Johansson, being highest with 20,195 votes, was declared suppléant to
Nordström.

This method of choosing the suppléant seems to be unsatisfactory. The party as such does not determine who shall be called upon to fill a vacancy in its ranks; whether a non-elected member succeeds to a vacancy as a suppléant depends very largely on accident. A good illustration occurred in the selection of a suppléant from the Labour list. The party's candidates were as follows:—

Kloo.
Karlsson.
Ostergren.
Olsson.
Ek.
Johansson.
Jensen.
Fagerberg.
Pettersson.

The first candidate on the list had been declared elected, and obviously, in the opinion of the party, the next favourite was Karlsson, and had there been a second seat awarded to the list Karlsson would have been declared elected. In determining, however, whether he should be declared elected as a suppléant, his position on the list did not count, and as the party list had been voted for without alteration by most of the Labour voters, five of the non-elected candidates were credited with the same number of votes. The choice of the suppléant was made by lot, and fell in this case upon Johansson, the sixth name on the list. It may be said that there is; considerable dissatisfaction with the method of electing suppléant candidates, and the Stockholm Dagblad, in its issue of the 29 May 1910, stated that the choice of suppléant, although there might have been many thousand votes given to every candidate, depended upon so small a difference in the totals received by each that even one ballot paper might determine the result. This is a detail in the system that can easily be remedied, and steps are already being taken to bring the election of suppléants into agreement with the election of ordinary members.

Comparison with Belgian system.

It will be of interest to compare the Swedish with the Belgian system. It has been shown that the method of allotting seats to different groups is identical in principle in both countries. This method, the d'Hondt rule, favours the largest parties, and this explains why, in the smaller Belgian constituencies, cartels or combinations of parties take place. The Swedish system enables such combined action to take place with greater facility. It enables two parties to make use of the same motto without presenting a common list of candidates. No inter-party negotiations are required, as in Belgium, with reference to the order in which the names of candidates shall appear upon the list. In Sweden each group can put forward its own list of candidates, and so long as the electors make use of the same motto at the head of the ballot paper the combination gains the additional representation which may fall to it as a result of being treated as one party, whilst the share falling to each section is determined by the number of votes recorded for their respective candidates.

The Swedish method of choosing the successful candidates from the various lists differs materially from that used in Belgium. In Sweden the d'Hondt rule is used not only for the allotment of seats to parties, but also in the selection of the successful candidates. In Belgium the use of the d'Hondt rule is restricted to the former purpose, and when once the electoral quotient is ascertained the rule is discarded. The difference in the two methods can be illustrated from the Stockholm municipal election of 1910. In the fifth ward the ballot paper of the Moderate party was as follows:—

Welin.
Norstrom.
Boalt.
Roberg.
Palmgren.
Bohman.
Ringholm.
Herlitz.
—————————
Hafstrom.
Svensson.
von Rosen.
Freden.

The line in the ballot paper divides the eight candidates for election as members from those who were standing for election as suppléants only. The votes recorded for the Moderate party numbered 118,483, of which 86,851 were given for the party ticket as printed. The number of votes accepting the party order of the first three candidates was about 93,000. This latter number was more than three-fourths, but less than four-fifths of the total, and therefore only the first three candidates on the ballot paper could be declared elected in accordance with the rule of the order of preference. The remaining four members had to be chosen by the reduction rule; the votes recorded for the five non-elected candidates were ascertained, the papers containing the names of the three elected candidates being treated for this purpose as of the value of one-fourth.

Some of the supporters of the eighth and sixth candidates had struck out the names of the fourth and other candidates. This manoeuvre had the result of placing these two candidates in the order named at the head of the poll at the fourth and fifth counts, and they were accordingly elected. Other candidates had received exclusive support, and it should be pointed out that it is the total amount of exclusive support recorded for all candidates which determines how soon the application of the rule of the order of preference breaks down. As soon as this takes place the election of any one candidate may depend, as in the election of the suppléants, upon the action of a comparatively small number of voters. Thus, some supporters of the fifth candidate, a Miss Palmgren, had struck out the names of all candidates save hers. Those papers which contained her name alone were treated as of full value, and although the votes of these supporters only numbered 1100, or less than 1 per cent. of the whole, they were sufficient to turn the scale in her favour. As, however, 86,851 votes out of a total of 118,453, had been recorded for the list as printed, showing that this proportion of voters preferred the fourth candidate to those that succeeded him, it would certainly seem that the result was not fair to this candidate. In Belgium if seven seats were won by a party which polled 118,453 votes, the electoral quotient would not be more than one-seventh of this total, and the election of the first candidate, instead of absorbing one-half the value of the votes, would consume only one-seventh. The election of the first two candidates would absorb two-sevenths instead of two-thirds, the election of three candidates would consume three-sevenths instead of three-fourths, and the election of four candidates would consume four-sevenths instead of four-fifths. In the Stockholm election more than five-sevenths of the voters had supported the party list as it was printed, and according to the Belgian system the first five candidates would have been declared elected.

The system and party organization.

The Swedish rule of selecting successful candidates is defended on the ground that it confers great power upon the electors. These can if necessary more effectively express their disapproval of the list put forward by the party organization, and as it is thought that a large number of voters too readily accept the party lead, a counterpoise is considered desirable. Recent experience in Belgium, however, would tend to show that a greater knowledge of their power has induced more and more electors to make use of the opportunity which that system allows of expressing individual preferences. If we regard a party as consisting of two groups—those that follow the party lead, and those which, whilst supporting the party, desire to assert their own preferences—then as between these two groups the Belgian system is strictly fair. If a party wins seven seats and four-sevenths of the party support the official list, this group would obtain four out of the seven seats; but in Sweden, as has been shown, at least four-fifths must support the official list before the first four candidates can be sure of election. The Swedish system discriminates in favour of the dissentients within a party, and this discrimination may have unexpected effects on party organization. The Belgian method has induced parties to welcome the support of all sections, knowing that such sections will not obtain more than their fair share of influence. In Sweden the tendency may be for party organizers to regard the support of various sections with suspicion, because, whilst these sections will obtain the full advantage of the party vote, their independent action may result in the gain of the section at the expense of the party as a whole. As a result of the Stockholm election referred to, the opinion was expressed by party organizers that it would be necessary to limit the number of candidates on a list to the number which the party knew it could carry. This would be an undesirable outcome of a rule designed to secure greater freedom for the elector, for it would tend to make party discipline more strict and parties exclusive rather than inclusive, as is the case in Belgium. It should, however, be added that in the large majority of the provincial council elections the selection of candidates was made in accordance with the rule of the order of preference. It would, therefore, seem that party organizers, as a rule, took care to present lists of candidates acceptable to the party as a whole.

The great improvement effected by the Swedish system.

The new Swedish electoral system, like all proportional systems, constitutes a striking advance upon the previous electoral conditions. The extent of the improvement will, of course, be seen from a comparison of some of its results with those of former years. For example, Stockholm used to be represented in the Lower Chamber by twenty-two members chosen by the "block" system, or scrutin de liste. The party in the majority monopolized the representation, and the absurdity of the system was well illustrated by an incident in the election of 1882, which was preceded by a severe struggle between the advocates of free trade and protection. At this election Stockholm returned twenty-two free traders, but as one of the elected members had not paid his taxes, all the voting papers containing his name were declared to be invalid. In consequence the twenty-two free traders were unseated and the twenty-two protectionist candidates were declared elected in their place. An attempt was made to ameliorate the evils of this system by dividing the town into five parliamentary districts, but, although so divided, Stockholm in 1908 returned twenty-one members, all of whom were either Liberals or Socialists, the large minority of Moderates being unrepresented. When the proportional system was applied in March 1910 to the election of the municipal council, each party obtained its fair share of representation in each of the six wards of the city, and the total result shows how large an improvement is effected by the new method:—

Parties. Votes Seats Seats in
Obtained. Obtained. Proportion
to Votes.
Moderate 281,743 22 24
Liberal 142,639 12 12
Socialist 160,607 16 14
—————————————————-
584,989 50 50

In the election of the provincial council of Bleking the result was as follows:—

Parties. Votes Seats Seats in Obtained. Obtained. Proportion to Votes. ———————————— ———————————- Moderate 54,465 22 22.4 Liberal 36,595 10 15.1 Socialist 3,617 1 1.5 ————————————————— 94,677 39 39

The general fairness of these results is all the more remarkable, because in Stockholm there was a very considerable variation in the value of a vote in the different wards, whilst many of the constituencies in the province of Bleking returned only a few members, and these did not give full play to the proportional system. The figures confirm the experience of all other countries, that a proportional system, even when applied to comparatively small constituencies, yields results which approximate very closely to the ideal aimed at, the true representation of the electors.

[Footnote 1: The town councils were elected in one stage; each elector had one vote for every 100 kroner income, subject to a limit of 100 votes. The members of the town council, when electing members of the provincial councils, had only one vote each.]

[Footnote 2: A ballot paper is not declared invalid even if it contains the names of more candidates than there are members to be elected (except at the elections of parliamentary committees). The names in excess are regarded as suppléant candidates (see Election of Suppléants) to the number of two in the elections for the Riksdag and the town councils, and to a number equal to the number of members at the election for the provincial councils. Any additional names on a ballot paper are regarded as non-existent.]

[Footnote 3: This paper bore the signature of the elector.]

APPENDIX IV

THE FINLAND SYSTEM OF PROPORTIONAL REPRESENTATION

The influence of the Belgian system.

The system of proportional representation introduced into Finland by the electoral law of 1906, while it presents little or no difficulty to the voter, is, in its method of counting the votes, perhaps the most complicated of the systems at present in force. It has for its basis the Belgian List system and the d'Hondt rule, but the variations which were introduced with the object of safeguarding the rights of the electors against the possible tyranny of party managers are so important that at the first glance its resemblance to the parent system is not easily recognized. The Belgian model is followed more closely in the method of distributing the seats to the various parties than in the manner in which the successful candidates are chosen from the party lists. In its internal party arrangement the Finnish system shows boldness, originality, and, it must be added, no little complexity of procedure.

Schedules and "compacts" in place of lists.

Finland is divided into sixteen electoral districts returning from six to twenty-three members, with the one exception of Lapland, which is a single-member constituency. In each constituency any group of not less than fifty electors can put forward a schedule of not more than three candidates, however many may be the total number of members to be elected. Each of these schedules may be headed with the name of a party or some political motto. The persons responsible for these schedules may, and commonly do, combine them in groups known as "compacts," and it is these compacts, and not the original schedules, which correspond roughly to the party "lists" of the Belgian system, the only limit to this power of combination being that the combined schedules must not contain the names of more candidates than there are vacancies to be filled. But as the names of the same candidates may, and constantly do, occur in many different schedules within a single compact, a first glance at a Finnish polling paper would seem to show in each combination the names of more candidates than there are vacancies. The compact bears the name of the political party to which it belongs. Combination into compacts is, of course, optional, and a certain number of schedules are put forward independently. A vacant corner is reserved on the ballot paper where any elector who is not content with any of the schedules submitted may make his own schedule.

An election in Nyland.

The system may be more fully understood from some details of the election of 1907 in the Nyland division. In this division, the largest in Finland, returning twenty-three members, no less than seventy-two schedules were presented, or which all except five were combined into compacts. The five remained isolated. Of the combined schedules seventeen were included in the compact of the Swedish party, but the individual candidates in these seventeen schedules numbered only twenty-three, the legal limit, the same names being repeated in several schedules. The old Finnish compact contained thirteen schedules, the Young Finns seventeen, the Social Democrats eight, the "Christian" compact seven, the "Free Christian" compact three, and the Radicals two.

As already stated, the voter's task is not difficult. He, or she, simply marks the schedule of his, or her, choice. The voter can also, if he wishes, alter the order of the names in a schedule. The effect of doing this will be apparent in a moment. That the task is simple is conclusively shown by the fact that the percentage of spoilt votes was in the Nyland division only 0.58 per cent. For the whole country the percentage was only 0.93, and this with universal adult suffrage and a poll of 899,347, or 70.7 per cent, of the electorate.

The returning officer's task.

The task of the returning officer is twofold. He has to ascertain (1) the relative positions of candidates within each compact (or independent schedule), and (2) their position relatively to the candidates of other compacts in the final allotment of seats. He proceeds as follows. He first counts the votes on each schedule, reckoning a full vote to the first name, a half vote to the second, and a third of a vote to the third (the effect of an alteration of the order of names in a schedule by the voter is now apparent). Thus if schedule No. 1 (in the specimen ballot paper on page 323), containing the names Schybergson, Neovius, and Soderholm, receives the support of 6000 voters in all, of whom 3000 have placed Schybergson as No. 1, 2000 as No. 2, and 1000 as No. 3, Schybergson will have a total of 3000 + 2000/2 + 1000/3 = 4333. Similarly, if Neovius obtains the support of 2000 as No. 1, 2000 as No. 2, and 2000 as No. 3, his total will be 2000 + 2000/2 + 2000/3 = 3666; Soderholm, the third candidate, would receive 1000 votes as No. 1, 2000 as No. 2, and 3000 as No. 3, and his total would be 1000 + 2000/2 + 3000/3 = 3000. But these individual totals of 4333, 3666, and 3000 are used merely to determine the order of the candidates within the schedule itself, and having performed that function, they are not taken further into account. In the example given (as would usually be the case in practice) the order within the schedule has not been disturbed, and the candidates are credited, the first (Schybergson) with the full number of the voters who supported the schedule—6000; the second (Neovius) with one-half that number—3000; the third (Soderholm) with one-third of that number—2000. These last figures are called "numbers of comparison," a phrase intended to throw light upon their function. The same process is gone through with all the other schedules in the same compact. The returning officer then adds up all the numbers of comparison which each candidate has obtained in all the schedules within the compact where his name appears, and arranges candidates within the compact in the order of these totals. Thus, in the actual election of 1907, in the Nyland division, Schybergson headed the Swedish party compact with 9192 as the total of his "numbers of comparison," Soderholm coming next with 6837.

The allotment of seats.

When the candidates in each compact have thus been arranged in order (and the votes given in writing by independent voters have also been counted), the returning officer proceeds to the second stage of his duties—the determination of the position of candidates with reference to their competitors in other compacts; and it is on this position that the actual allotment of seats depends. For this purpose he primarily takes into account, not the "numbers of comparison" of individual candidates, but the total number of voters who have supported each compact; he credits this total to the candidate who has the highest "number of comparison" within the compact; credits the next candidate with one-half this total, the third candidate with one-third, and so on, finally arranging the whole of the candidates in order. Thus far this stage of the process is identical in substance with the Belgian method, though the appearance is different. For, obviously, if List (or compact) A, of which the candidates are G, H, I, in that order receives 12,000 votes, while List B, with candidates P, Q, R, receives 10,000, and List C, with candidates X, Y, Z, receives 8000, it is all one whether the returning officer applies the d'Hondt rule and assigns two seats to List A (thus seating G and H), two seats to List B (thus seating P and Q), and one seat to List C (thus seating X), or whether he tabulates the result of the polling thus:

G 12,000 \
P 10,000 |
X 8,000 > Elected.
H 12,000/2 i.e. 6,000 |
Q 10,000/2 i.e. 5,000 /
Y 8,000/2 i.e. 4,000 Not elected, and so on.

But at this point a characteristic feature of the Finnish system comes into play. Candidates' names may occur in more than one compact, and may be found in isolated schedules, or on the written papers of independent voters as well. Consequently their final order cannot be determined by this simple application of the Belgian method. The returning officer must[1] add to the number of votes credited to a candidate of any one compact such additional votes as he may have obtained either as a member of another compact or from independent voters. Thus, in the Nyland elections, Miss Sohlberg, whose name will be found at the head of Schedule 48 within the Swedish compact, obtained the eleventh place within that compact. The total number of voters supporting this compact was 44,544, and Miss Sohlberg was therefore credited with an eleventh of this total, or 4049 votes. But Miss Sohlberg's name also occurred in Schedules 62 and 63 in the "Free Christian" compact and Schedule 21 in the "Christian" compact, and as her share of the votes of these compacts she received 153 and 325 respectively. She also received four votes in writing. Thus her final total was 4049 + 153 + 325 + 4, or 4531 in all, and it was this number which determined her position on the poll.

Successful candidates in the Nyland election. This explanation will perhaps be more comprehensible if the actual result of the polling in the Nyland division, so far as the first 25 candidates are concerned, is given in a tabular form:—

Final Names of Party. Number of Additional Final Order Candidates. Votes resulting Votes. Total. of from Place of Poll. Candidates on Compact. 1 Schybergson Swedish 44,544 2.33 44,546.33 2 Häninan Social Dem. 40,951 6.5 40,957.5 3 Soderholm Swedish 22,272 0.33 22,272.33 4 Sillanpää Social Dem. 20,475.5 8.83 20,484.33 5 Käkikoski Old Finn 20,402 9.33 20,411.33 6 Oljemark Swedish 14,848 — 14,848 7 Sirén Social Dem. 16,650.33 2.33 16,652.66 8 Rosenquist (G.) Swedish 8,908.8 2,932.83[2] 11,841.63 9 Rosenquist (V.) Swedish 11,136 4.33 11,140.33 10 Helle Social Dem. 10,237.75 3 10,240.75 11 Palmén Old Finn 10,201 8.83 10,209.83 12 Pertillä (E.) Social Dem. 8,190.2 4.67 8,194.87 13 Ahlroos Swedish 7,424 1 7,425 14 Pertillä (V.) Social Dem. 6,725.17 1.5 6,726.67 15 Reima Old Finn 6,800.67 5.67 6,806.34 16 Erkko Young Finn 6,521 6.32 6,527.32 17 Ehrnrooth Swedish 6,363.43 75.83 6,439.26 18 Laine (M.) Social Dem. 5,850.14 4 5,854.14 19 Wasastjerna Swedish 5,568 — 5,568 20 Ingman Social Dem. 5,118.88 3.5 5,122.38 21 Laine (O.) Old Finn 5,100.5 — 5,100.5 22 von Alfthan Swedish 4,949.33 — 4,949.33 23 Johansson Social Dem. 4,550.11 1.33 4,551.44 (All the above were elected.) 24 Sohlberg Swedish 4,049.45 482.45[3] 4,531.9 25 Gustaffsson Swedish 4,454.4 4.5 4,458.9 &c. &c.

Equitable results.

It will to some extent be gathered from the foregoing table that the total number of the supporters of the various compacts or parties in the Nyland division and the number of seats won were as follows:

Seats Seats in
Parties. Votes. Actually Proportion
Won. to Votes.
Swedish 44,544 9 8.7
Social Democrat 40,951 9 8.0
Old Finn 20,402 4 4.0
Young Finn 6,521 1 1.3
"Christian" compact 2,932 - .6
"Free Christian" 458 - .1
Radical 168 - -
Isolated schedules 1,356 - .3

Total 117,332 23 23.0

The result is thus in reasonable correspondence with the demands of a strictly proportionate allotment of seats; this statement is also true of the results for the whole of Finland, as the following table will show:—

Seats Seats in
Parties. Votes. Actually Proportion
Won. to Votes.
Social Democrat 329,946 80 74.1
Old Finn. 243,573 59 54.7
Young Finn 121,604 26 27.3
Swedish 112,267 24 25.2
Agrarian 51,242 9 11.5
Christian Labourer 13,790 2 3.1
Minor groups 18,568 - 4.1

Total 890,990 200 200.0

An exactly mathematical distribution is, of course, not to be expected from this, any more than from any other method which does not adopt the system of treating a whole country as a single constituency. As to the mechanism of the system it only remains to add that the process of counting was found to be very lengthy. In the Nyland division, where the results were ascertained sooner than in any other case, the elections were held on 15 and 16 March, but the result was not announced until the 2 April. To people accustomed to the greater rapidity of ordinary electoral methods this will seem a serious drawback. Possibly improved arrangements may shorten this long interval between the elections and the announcement of the result.

It would obviously be premature to attempt to estimate the political effects of the Finnish system as compared with other systems of proportional representation.

Elector's freedom of choice.

The Finnish system has been in operation since 1907, and the whole political circumstances of Finland have undergone so many striking changes, and so many new factors are at work that to disentangle particular causes and effects is an impossibility. But plainly the Finnish machinery gives a greater freedom to the elector than the Belgian system. The Finnish system in fact encourages the electors to arrange the candidates of a party in the order preferred by the electors themselves, and not in the order dictated by the party managers. There is no "party ticket" for which the elector can vote blindfold. He must choose the schedule that he prefers; he can even rearrange that schedule, or, if he chooses, can make one of his own. No doubt the schedule itself is ready made for him, but it contains three names only, and is not the equivalent of the Belgian "list." On the other hand, the elector who chooses to vote for a schedule within a compact adds, whether he likes it or not, to the total votes of the compact, and so may help to return not the candidate of his choice, but the candidates preferred by the majority of the party with which he is in sympathy. An illustration of this fact may be taken from the Nyland poll. The old Finnish party were alive to the possibilities of the situation, and combined their lists with great skill so as to attract votes. They placed their favourite candidates in nearly every schedule, but not at the head of the schedule. At the head of the schedule they placed some man of local popularity, usually a peasant proprietor, whose name was not repeated in many, if any, other schedules. Thus the local favourite attracted votes to the schedule, but in the race for the highest numbers of comparison the candidates whose names appeared on few schedules were left behind those whose names appeared on many schedules even in the lower places.

A portion of the official ballot paper showing the compact put forward by the Swedish People's Party is printed on the opposite page. In one corner of the ballot paper was a blank schedule in the following form.

THE ELECTOR who does not approve of any of the preceding lists should write here the names of his candidates in the order in which he wishes them to be elected.

CANDIDATES

Name…………………………………………….

Profession or Occupation…………………………..

Address………………………………………….

Name…………………………………………….

Profession or Occupation…………………………..

Address………………………………………….

Name…………………………………………….

Profession or Occupation…………………………..

Address………………………………………….

FINLAND GENERAL ELECTION, 1907

Part of Ballot Paper—Nyland Division.

The Voters' Compact of the Swedish People's Party.

1
HELSINGFORS.
Experienced Members of the Diet:—
—Schybergson, E. K.
—Neovius, A. W.
—Soderholm, K. G.

33
EAST NYLAND-LOUISA.
Justice and Progress:—
—Rosenquist, G. G.
—Stromberg, J.
—Ehrnrooth, L.

34
MID-NYLAND-NIOKBY.
The Welfare of the Rural Population;—
—Topelius, G. L.
—Alfthau, K. von
—Rosenquist, B. T.

35
MID-NYLAND-ESBO.
The Welfare of the Rural Population:—
—Wasastjerna, O.
—Schybergson, E.
—Soderholin, K.

36
WEST NYLAND-KYRK-SLATT.
The Welfare ol the Rural Population:—
—Nordberg, G.
—Ehrnrooth, L.
—Oljemark, K. T.

37
WEST NYLANB-EKENAS.
The Welfare of the Rural Population. Law and Justice:—
—Oljemark, K. T.
—Schybergson, E.
—Soderholm, K.

38
BORGA.
Knowledge and Experience:—
—Runeberg, J. W.
—Bjorkenheim, G.
—Rosenquist, G. G.

39
HELSINGFORS.
Sound Development of the Community;—
—Westermarck, Helena.
—Rosenquist, B. T.
—Bjorkenheim, G.

40
HELSINGFORS.
Law and Justice:—
—Sorterholm, K.
—Alfthan, K. von
—Westermarck, Helena,

41
HELSINGFORS.
Legality and Progress:—
—Westermarck, Helena.
—Neovius, A.
—Ehrnrooth, L.

42
HELLSINGFORS.
Swedish Culture:—
—Rosenqnist, B. T.
—Gustafsson, F. prof.
—Soderholm, K.

43
HELSINGFORS.
Friends of Labour and of the People:—
—Alfthan, K. von
—Gustafsson, F. prof.
—Gronroos, F.

44
HELSINGFORS.
Experience and Practical Knowledge:—
—Runeberg, J. W.
—Schybergson, E.
—Neovius, A.

45
HELSINGFORS.
The Labourers' Welfare:—
—Ahlroos, F.
—Holmberg, W.
—Ehrnrooth, L.

46
HELSINGFORS.
Commerce and Industry:
—Heimburger, W. F.
—Bjorkenheim, G.
—Schybergson, E.

47
THE SKERRIES OF NYLAND:
Navigation and Fisheries:—
—Hjelt, Th.
—Renter, O.
—Alfthan, K.

48
THE PROVINCE OF NYLAND:
HELSINGFORS.
Temperance, Morality and Popular Education:—
—Sohlberg, H.
—Ahlroos, F.
—Rosenquist, G. G.

[Footnote 1: This right of addition is subject to a limit. The reinforcements must not raise a candidate's total above what he might obtain if the votes given to all compacts or lists, where his name occurs, were divided by the figure which indicates his order within the compact from which he derives his principal strength.]

[Footnote 2: This large reinforcement of votes came from the Christian compact, where this candidate's name appeared as well as in the Swedish compact.]

[Footnote 3: See reference to Miss Sohlberg in preceding paragraph.]

APPENDIX V

THE STATISTICS OF THE GENERAL ELECTIONS, 1885-1910

The following tables are taken, with permission, from a paper read on 12
December 1906, by Mr. J. Rooke Corbett, M.A., before the Manchester
Statistical Society, of which a second and revised edition was published
in April 1910 by the Proportional Representation Society.

In these tables the totals for England, Wales, and Monmouth, Scotland and Ireland are shown separately, and the figures for England have been further subdivided according to the ten divisions into which the kingdom is divided by the Registrar General for the purpose of his work.

These ten subdivisions are as follows:

Metropolitan—
London.
South East—
Surrey.
Kent.
Sussex.
Hampshire.
Berkshire.
South Midland—
Middlesex.
Hertfordshire.
Buckinghamshire.
Oxfordshire.
Northamptonshire.
Huntingdonshire.
Bedfordshire.
Cambridgeshire.
East—
Essex.
Suffolk.
Norfolk.
South-West—
Wiltshire.
Dorsetshire.
Devonshire.
Cornwall.
Somersetshire.
West Midland—
Gloucestershire.
Herefordshire.
Shropshire.
Staffordshire.
Worcestershire.
Warwickshire.
North Midland—
Leicestershire.
Rutlandshire.
Lincolnshire.
Nottinghamshire.
Derbyshire.
North-West—
Cheshire.
Lancashire.
Yorkshire—
West Riding.
East Riding (with York).
North Riding.
Northern Division—
Durham.
Northumberland.
Cumberland.
Westmorland.

The first three columns, A, B and C, show the number of members allotted to these several divisions, the number of registered electors, and the number of members to which each division would be entitled if the 670 members of which the House of Commons is composed were divided among the several divisions in proportion to their electorates.

In taking the electorate as the basis of a proportionate redistribution of seats it is not intended to prejudge the question whether population or electorate is the better standard. The electorate has been taken because the figures are available for the very year in which the election takes place, whereas the population is only enumerated once in ten years.

The columns D and E show in two groups the number of members elected for these divisions, Liberal, Labour, and Irish members being gathered together in one column, Conservatives alone occupying the other.

It is one of the disadvantages of our present system of representation that it makes it quite impossible to ascertain the relative strength of the several parties into which the voters are divided. In the great majority of contests there is a Liberal, Labour, or Irish Nationalist candidate on one side, and a Unionist candidate on the other, and there is practically no evidence as to how many of the supporters of either candidate belong to each of the parties concerned. Any estimate of the relative strength of the Liberal and Labour parties or of the Unionist Free Traders, and Tariff Reformers must be largely a matter of guesswork. All that is possible, therefore, is to divide the voters into two groups, as has been done in these tables.

The columns F and G show the total electorate of the constituencies held respectively by the two groups of members shown in columns D and E.

The figures in these two columns are of value in showing the probable result of a scheme of redistribution. The South-Eastern counties may be taken as an example. These are at present represented by 48 members. The Liberals held three constituencies in January 1910 containing an electorate of 31,221 (columns D and F); the Conservatives held 45 constituencies containing an electorate of 604,887 (columns E and G). If a redistribution of seats was made on the basis of equal electorates, the South-Eastern counties would be entitled to 55 members (column C). It may be assumed that in any rearrangement of constituencies the parties would retain their predominance in the areas which they now represent, and if so the result of a rearrangement of constituencies on the basis of equal electorates would be that in January 1910 the Conservatives would have obtained 52 seats and the Liberals 3 (column K). Similarly in the General Election of 1906 the Liberals in Wales and Monmouth held 34 seats, the Conservatives none. If the constituencies had been rearranged, the Liberals would have held 35 seats, the Conservatives none. The majorities throughout the United Kingdom which would be obtained under a scheme of equal electorates are shown in column K.

The columns H and I show the number of electors who voted for the candidates of the two groups; Liberal, Labour, and Irish Nationalist voters in one group, Conservative voters in the other.

In computing the figures in these columns an allowance has been made for uncontested constituencies on the following basis. It has been supposed that the changes of public opinion which affect the contested constituencies affect uncontested constituencies also, and in estimating the number of voters in an uncontested constituency it has therefore been assumed that the strength of each party varies from one election to another in the same ratio as in the contested constituencies in the same county.

The three columns J, K and L show respectively the actual majorities obtained, the majorities which would have been obtained if the country had been divided into single-member constituencies of equal size, and the majorities under a system of proportional representation.

The figures in the last two columns have been calculated with reference to the totals in column C, which gives the number of members to which each division would be entitled on a proportional basis.

In order to ascertain the figures given in column K (i.e. the probable results with equal single-member constituencies) it has been assumed, as already explained, that the two groups would, after the redistribution of seats, be predominant in the same areas as before the rearrangement.

The representation of minorities.

The tables give abundant evidence of the anomalies associated with our electoral system. One of the most striking is the great difference in the amount of representation secured by minorities in different parts of the country. The amount of representation secured by a minority has not depended upon its size, but upon the way in which it has been distributed. The following table shows the amount of representation obtained by important minorities in the General Election of January 1910:—

THE REPRESENTATION OF MINORITIES, ELECTION JAN. 1910

Size of Seats Total Seats
Area. Minority. Obtained. for Whole Area
Ireland . . . . . . . 145,437 21 103
Scotland . . . . . . . 265,770 11 72
S. East: Counties. . . 220,995 3 48
Wales and Monmouth . . 116,696 2 34
Northern Counties . . 75,897 9 32

The figures show that in Ireland a minority of 145,437 obtained twenty-one representatives, whilst a minority of 116,696 in Wales and Monmouth obtained only two. The good fortune which befel the minority in Ireland, not only in the elections of 1910 but in all the elections since the Redistribution Bill of 1885, has been due to the fact that this minority is concentrated in one corner of Ireland and can transform itself into local majorities. The larger minority in Scotland, owing to its distribution throughout the country, obtains much less representation; the minorities in the south-eastern counties of England and Wales are also distributed throughout these two areas and likewise suffer. The minority of 75,879 in the northern counties being less evenly diffused was more fortunate, and obtained nine representatives. The figures for the election of December 1910 disclose similar anomalies.

GENERAL ELECTION, 1885

Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 489,396 57 LLI 22 165,345 162,228
Con 38 324,051 188,067 16 19 3
England
South-East 48 406,955 47 LLI 4 34,883 144,659
Con 44 372,072 187,831 40 39 7
S.Midland 38 312,477 36 LLI 14 123,665 124,717
Con 24 188,811 129,544 10 8
East 29 257,022 29 LLI 18 173,521 107,710 7 11 1
Con 11 83,501 98,137
South-West 40 314,603 36 LLI 27 229,612 144,273 14 16 4
Con 13 84,991 117,442
W.Midland 58 544,415 63 LLI 45 427,549 248,825 32 36 8
Con 13 116,866 198,212
N.Midland 34 328,844 38 LLI 26 255,836 55,503 18 22 4
Con 8 73,008 120,933
North-West 70 654,751 76 LLI 24 231,123 263,670
Con 46 423,628 292,942 22 22 4
Yorkshire 52 536,553 62 LLI 36 398,426 248,078 20 30 8
Con 16 138,127 189,930 20 30 8
North 32 305,015 35 LLI 25 262,287 144,803 18 25 5
Con 7 42,728 96,708
ENGLAND 461 4,150,031 480 LLI 241 2,302,248 1,740,466 21 52 16
Con 220 1,847,783 1,619,746
Wales and
Monmouth 34 286,145 33 LLI 30 263,199 149,782 26 27 11
Con 4 22,946 79,006
Scotland 72 576,828 67 LLI 58 485,116 289,032 44 45 15
Con 14 91,712 181,706

Britain 567 5,013,004 580 LLI 329 3,050,563 2,179,230 91 124 42
Con 238 1,962,441 1,880,458
Ireland 103 777,954 90 LLI 85 624,760 404,892 67 54 44
Con 18 153,194 139,273

Total 670 5,790,958 670 LLI 414 3,675,323 2,584,122 158 178 86
Con 256 2,115,635 2,019,731

Majority 158 1,559,638 564,391

NOTE.—The figures in columns K and L are calculated with reference to the totals in column C. Thus the figure L 54 for Ireland in column K of the last section of the table indicates that under a system of equal single-member constituencies Ireland's 90 members would be Liberal etc. 72, Unionist 18, a Liberal majority of 54, and the corresponding figure L 44 in column L indicates that under proportional representation the 90 members which Ireland would return would be Liberal etc. 67, and Unionist 23. a Liberal majority of 44.

GENERAL ELECTION, 1886

Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 489,396 57 LLI 11 87,974 125,457
Con 49 401,422 185,072 38 37 11
England—
South-East 48 406,955 47 LLI 0 - 114,518
Con 48 406,955 184,221 48 47 11
S.Midland 38 312,477 36 LLI 9 73,292 94,213
Con 29 239,185 128,339 20 20 6
East 29 257,022 29 LLI 4 87,975 81,838
Con 25 219,047 102,732 21 21 3
South-West 40 314,603 36 LLI 7 63,063 96,753
Con 33 251,540 129,056 26 22 6
W.Midland 58 544,415 63 LLI 15 136,518 173,463
Con 43 407,897 218,753 28 32 8
N.Midland 34 328,844 38 LLI 14 147,138 125,078
Con 20 181,706 126,547 6 4
North-West 70 654,751 76 LLI 13 123,459 236,134
Con 57 531,292 282,187 44 48 6
Yorkshire 52 536,553 62 LLI 33 359,414 214,407 6
Con 19 177,139 180,728 14 22
North 32 305,015 35 LLI 23 247,275 123,901 5
Con 9 57,740 96,404 14 21
ENGLAND 461 4,150,031 480 LLI 129 1,276,108 1,385,762
Con 332 2,873,923 1,634,039 203 188 42

Wales and
Monmouth 34 286,145 33 LLI 27 240,752 123,186 20 23 7
Con 7 45,393 82,179
Scotland 72 576,828 67 LLI 43 339,726 218,561 14 11 5
Con 29 237,102 188,164

Subtotal 567 5,013,004 580 LLI 199 1,856,586 1,727,509
Con 368 3,156,418 1,904,382 169 154 30

Ireland 103 777,954 90 LLI 84 616,735 376,445
Con 19 161,219 144,755 65 52 38

Total 670 5,790,958 670 LLI 283 2,473,321 2,103,954 8
Con 387 3,317,637 2,049,137 104 102

Majority 104 844,316 54,817

GENERAL ELECTION, 1892

Table headings:
Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 552,024 60 LLI 23 186,572 183,967
Con 37 365,452 214,275 14 20 4
England:
South-East 48 463,073 50 LLI 4 38,534 147,136
Con 44 424,539 206,075 40 42 8
S.Midland 38 340,650 38 LLI 15 139,228 120,844
Con 23 210,422 147,347 8 8 4
East 29 276,491 30 LLI 13 134,632 108,866
Con 16 141,859 110,849 3
South-West 40 325,769 35 LLI 15 136,061 125,392
Con 25 189,708 136,449 10 5 1
W. Midland 58 577,397 63 LLI 16 143,567 204,453
Con 42 433,830 248,774 26 31 7
N. Midland 34 347,482 38 LLI 22 232,970 145,587 10 14 2
Con 12 114,512 130,380
North-West 70 707,392 77 LLI 26 284,970 282,139
Con 44 422,422 307,698 18 15 3
Yorkshire 52 571,864 62 LLI 35 418,414 244,099 18 28 6
Con 17 153,450 204,492
North 32 328,189 36 LLI 25 264,483 143,172 18 22 4
Con 7 63,706 115,626
ENGLAND 461 4,499,331 489 LLI 194 1,979,431 1,705,655
Con 267 2,519,900 1,821,985 73 57 15

Wales and
Monmouth 34 314,063 34 LLI 31 294,395 152,326 28 30 10
Con 3 19,668 86,576
Scotland 72 606,203 66 LLI 52 449,994 267,631 32 32 8
Con 20 156,209 214,448

Subtotal 567 5,419,497 589 LLI 277 2,723,820 2,125,612 5 3
Con 290 2,695,777 2,123,009 13
Ireland 103 746,781 81 LLI 80 561,938 345,548 57 41 31
Con 23 184,843 157,181

Total 670 6,168,388 670 LLI 357 3,285,758 2,471,164 44 46 34
Con 313 2,880,620 2,280,190
Majority 44 405,138 190,974

GENERAL ELECTION, 1895

Table headings:
Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 573,141 61 LLI 8 70,056 161,328
Con 52 503,085 242,999 44 47 13
England:
South-East 48 472,725 50 LLI 2 24,057 152,213
Con 46 448,668 217,096 44 44 8
S.Midland 38 358,501 38 LLI 3 30,569 116,143
Con 35 327,932 164,052 32 32 6
East 29 294,153 31 LLI 8 70,467 101,736
Con 21 223,686 122,999 13 15 3
South-West 40 330,670 35 LLI 10 76,141 124,852
Con 30 254,529 144,435 20 19 3
W.Midland 58 589,881 63 LLI 9 85,544 195,545
Con 49 504,337 259,382 40 45 9
N.Midland 34 351,792 37 LLI 16 186,167 143,142 1
Con 18 165,625 149,436 2 1
North-West 70 728,292 78 LLI 10 114,035 273,585
Con 60 614,257 332,101 50 54 8
Yorkshire 52 565,799 61 LLI 28 317,932 238,032 4 7 1
Con 24 247,867 225,871
North 32 339,289 36 LLI 20 222,202 145,085 8 12 2
Con 12 117,087 124,697

ENGLAND 461 4,604,243 490 LLI 114 1,197,170 1,652,261
Con 347 3,407,073 1,983,068 233 236 48
Wales and
Monmouth 34 320,532 34 LLI 25 241,750 148,552 16 18 6
Con 9 78,782 108,036
Scotland 72 636,106 68 LLI 39 335,143 243,425 6 4 2
Con 33 300,963 234,138

Subtotal 567 5,560,881 592 LLI 178 1,774,068 2,044,238
Con 389 3,786,818 2,325,242 211 214 40

Ireland 103 727,562 78 LLI 82 549,467 317,910 61 42 28
Con 21 178,095 154,379

Total 670 6,292,443 670 LLI 260 2,323,530 2,362,148
Con 410 3,964,913 2,479,621 150 172 12
Majority 150 1,641,383 117,473

GENERAL ELECTION, 1900
Table headings:
Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 601,925 60 LLI 8 73,718 150,047
Con 52 528,207 247,777 44 46 14
England:
South-East 48 512,408 51 LLI 3 23,362 140,277
Con 45 489,406 220,829 42 47 11
S. Midland 38 388,361 39 LLI 6 63,375 120,012
Con 32 324,986 164,148 26 27 7
East 29 319,997 32 LLI 9 80,447 101,785
Con 20 239,550 125,375 11 8 4
South-West 40 337,449 33 LLI 14 122,410 127,086
Con 26 215,039 142,269 12 9 1
W. Midland 58 630,931 63 LLI 10 96,089 200,113
Con 48 534,842 261,474 38 43 9
N. Midland 34 378,996 38 LLI 18 211,280 149,794 2 4 0
Con 16 167,716 153,294
North-West 70 794,142 79 LLI 14 176,183 281,634
Con 56 617,957 351,243 42 43 9
Yorkshire 52 612,892 61 LLI 26 326,841 239,045 5 1
Con 26 286,051 238,870
North 32 367,007 36 LLI 16 197,102 147,017 2 2
Con 16 169,905 135,459
ENGLAND 461 4,944,108 492 LLI 124 1,370,807 1,657,814
Con 337 3,573,301 2,040,508 213 212 52

Wales and
Monmouth 34 342,209 34 LLI 28 286,628 161,190 22 24 8
Con 6 55,581 103,396
Scotland 72 683,840 68 LLI 34 312,781 254,112
Con 34 371,059 258,836 4 6

Britain 567 5,970,187 594 LLI 186 1,970,216 2,073,116
Con 381 3,999,941 2,402,740 195 194 44

Ireland 103 765,258 76 LLI 82 598,469 318,203 61 44 28
Con 21 166,757 145,906

Total 670 6,735,415 670 LLI 268 2,568,685 2,391,319
Con 402 4,166,698 2,548,736 134 150 16
Majority 134 1,598,013 157,417

GENERAL ELECTION, 1906

Table headings:
Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 626,011 57 LLI 40 385,762 251,937
Con 20 240,249 225,725 20 13 3
England
South East 48 583,000 54 LLI 22 273,398 245,046
Con 26 309,602 241,097 4 4
S.Midlands 38 441,803 40 LLI 27 328,386 193,594 16 20 2
Con 11 113,417 172,159
East 29 368,662 34 LLI 25 333,564 170,039 21 28 4
Con 4 35,098 128,991
South-West 40 371,300 34 LLI 34 321,822 176,478 28 24 4
Con 6 49,478 144,342
W.Midland 58 679,903 63 LLI 35 402,148 288,832 12 11 1
Con 23 277,760 286,862
N.Midland 34 420,677 39 LLI 28 358,852 205,066 22 27 5
Con 6 61,825 151,924
North-West 70 869,792 80 LLI 55 680,843 420,969 40 46 12
Con 15 188,949 321,560
Yorkshire 52 667,863 62 LLI 41 556,233 340,865 30 42 14
Con 11 111,635 218,778
North 32 409,843 38 LLI 27 345,353 215,748 22 26 10
Con 5 64,490 123,003
England 461 5,438,859 501 LLI 334 3,986,356 2,508,574 207 233 53
Con 127 1,452,503 2,014,441

Wales and
Monmouth 34 387,585 35 LLI 34 387,585 217,462 34 35 13
Con 0 — 100,547
Scotland 72 750,401 70 LLI 60 629,360 367,942 48 48 16
Con 12 121,041 235,098

Britain 567 6,576,845 606 LLI 428 5,003,301 3,093,978 289 316 82
Con 139 1,573,544 2,350,086

Ireland 103 693,417 64 LLI 85 545,748 301,833 67 36 22
Con 18 147,669 144,708

TOTAL 670 7,270,262 670 LLI 513 5,549,049 3,395,811 356 352 104
Con 157 1,721,213 2,494,794
Majority 356 3,827,836 901,017

GENERAL ELECTION, JANUARY 1910

Table headings:
Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 658,795 57 LLI 26 246,838 254,154
Con 34 411,957 298,821 8 15 5
England:
South-East 48 636,108 55 LLI 3 31,221 220,995
Con 45 604,887 334,022 42 49 11
S. Midland 38 490,592 43 LLI 11 146,312 197,717
Con 27 344,280 235,776 16 17 3
East 29 400,062 35 LLI 15 236,234 173,465 1 7 1
Con 14 163,828 170,027
South-West 40 386,514 34 LLI 18 201,726 172,692 2
Con 22 184,788 175,010 4
W. Midland 58 713,761 62 LLI 17 227,430 284,629
Con 41 486,331 334,874 24 22 6
N. Midland 34 446,752 39 LLI 23 334,766 216,469 12 19 3
Con 11 111,986 181,209
North-West 70 928,640 81 LLI 47 636,497 449,324 24 35 7
Con 23 292,143 382,796
Yorkshire 52 701,856 61 LLI 89 564,418 365,185 26 37 11
Con 13 137,438 248,507
North 32 430,594 38 LLI 23 354,697 216,760 14 24 6
Con 9 75,897 150,471
ENGLAND 461 5,793,674 505 LLI 222 2,980.139 2,551,390 21 3
Con 239 2,813,535 2,521,513 17
Wales and
Monmouth 34 425,714 37 LLI 32 414,613 243,383 30 35 13
Con 2 11,101 116,696
Scotland 72 785,391 68 LLI 61 675,723 394,103 50 50 14
Con 11 109,668 265,770
Sub total 567 7,004,779 610 LLI 315 4,070,475 3,188,876 63 106 30
Con 252 3,188,876 2,903,979

Ireland 103 688,284 60 LLI 82 518,154 356,223 61 30 26
Con 21 170,130 145,437

Total 670 7,693,063 670 LLI 397 4,588,629 3,545,099 124 136 56
Con 270 3,104,434 3,049,416
Majority 124 1,484,195 495,683

GENERAL ELECTION, DECEMBER 1910

Table headings:
Col A: Members
Col B: Registered Electors
Col C: Proportionate Number of Members
Col D: Members - Liberal, Labour and Irish
Col E: Members - Conservatives
Col F: Electorate of Constituencies held by - Liberal, Labour,
and Irish Nationalists
Col G: Electorate of Constituencies held by - Conservative
Col H: Voters - Liberal, Labour, and Irish Nationalist
Col I: Voters - Conservative
Col J: Majority - Actual
Col K: Majority - With equal Single Member Constituencies
Col L: Majority - Under Proportional Representation.

A B C DE FG HI J K L
Prop Memb Electorate Voters Majority
Memb Elect Memb Act Eq PR
Metropolis 60 658,795 57 LLI 29 279,492 223,151
Con 31 379,303 264,281 2 9 5
England—
South-East 48 636,108 55 LLI 5 58,248 209,434
Con 43 577,860 311,888 38 45 11
S. Midland 38 490,592 43 LLI 14 170,762 190,120
Con 24 319,830 219,876 10 13 3
East 29 400,062 35 LLI 16 256,750 164,849 3 9 1
Con 13 143,312 154,529
South-West 40 386,514 34 LLI 14 159,494 164,698
Con 26 227,020 168,992 12 6 0
W. Midland 58 713,761 62 LLI 19 246,842 268,125
Con 39 466,919 316,574 20 20 6
N. Midland 34 446,752 39 LLI 21 298,037 202,351 8 13 3
Con 13 148,715 173,545
North-West 70 928,640 81 LLI 39 524,682 400,508 8 11 1
Con 31 403,958 386,045
Yorkshire 52 701,856 61 LLI 40 570,544 321,622 28 39 9
Con 12 131,312 239,067
North 32 430,594 38 LLI 25 375,574 200,583 18 28 6
Con 7 55,020 142,388
ENGLAND 461 5,793,674 505 LLI 222 2,940,425 2,345,441 7
Con 239 2,853,249 2,377,185 17 5
Wales and
Monmouth 34 425,714 37 LLI 31 388,507 210,525 28 31 9
Con 3 37,207 121,013
Scotland 72 785,391 68 LLI 61 678,395 372,313 50 50 10
Con 11 106,996 277,183

Subtotal 567 7,004,779 610 LLI 314 4,007,327 2,928,279 61 88 14
Con 253 2,997,452 2,775,381

Ireland 103 688,284 60 LLI 84 536,675 350,029 65 34 24
Con 19 151,609 146,982

Total 670 7,693,063 670 LLI 398 4,544,002 3,278,308 126 122 38
Con 272 3,149,061 2,922,363
Majority 126 1,394,941 355,945

APPENDIX VI

PREFERENTIAL VOTING: THE TRANSFER OF SUPERFLUOUS VOTES

(A Memorandum by the Rt. Hon. J. Parker Smith)[1]

(1) The Element of Chance Involved: Its Magnitude

An objection, which occurs to every one who considers schemes of Preferential Voting, is that an element of chance is introduced into the result by the methods for the transfer of the superfluous votes of successful candidates. Supposing one part of the supporters of A, a successful candidate, have put down B as their second choice, and the remainder C, and that a certain number of A's votes are superfluous, and have to be transferred, how is it to be determined what number of AB votes, as they may be called, and what number of AC votes shall be transferred? If the question is settled by chance, as, by drawing the necessary number at random from A's heap, by declaring that voting papers shall be used in the order in which they were handed in at the polling booths, or by laying down any other set of arbitrary rules to determine the order in which they shall be counted, an element of uncertainty is introduced by which there seems to be serious danger that B and C will gain or lose unfairly.

Those who are accustomed to dealing with statistics will be prepared to find this danger less than might have been expected; but even they will be surprised to find of how small importance the arbitrary element is discovered, by actual calculation, to be.

The difficulty can be made clear by a numerical instance. Take the case of an election for several seats, where the necessary quota is 6000, and where a favourite candidate, whom we will call A, has received the first votes of 10,000 voters. Though all those voters have agreed in putting the same candidate first, they are divided as to who may wish to be returned next. Six thousand of them put B as their second choice, and the other 4000 C. If the 6000 votes which A requires are drawn wholly from the AB votes, the result of the transfer will be that C is credited with 4000 votes and B with none. This would be clearly unfair, for, in reality, B has received among A's voters much more support than C. To use up the 4000 AC votes and only 2000 AB votes, and to transfer 4000 votes to B and none to C would be equally unfair to C. The course which is exactly fair to both B and C is that the votes which are transferred should be divided between them in the same proportion as that in which the opinions of the whole number of A's supporters is divided. That is to say, strict justice will be done if every 1000 votes which are used or transferred are made up of 600 AB votes and 400 AC votes. Accordingly, A's quota of 6000 must be made up of 3600 AB votes and 2400 AC votes, and the 4000 papers left to be transferred will consequently consist of 2400 votes for B and 1600 votes for C.

This principle avoids all uncertainty, and is indisputably fair. It remains to consider how to carry it into effect. In most cases there would, in reality, be many more classes of votes than in the instance taken above. Even in such cases it is practicable, as will presently be shown, to divide the votes proportionately by an actual process of counting and separation. A certain amount of complication is, of course, introduced, but the extra labour involved does not seem impossible. The question whether this extra labour is necessary must be answered by examining the magnitude of the evil which it is sought to remedy.

If the votes are counted in a random order, it is clear there is a probability that the order in which they are drawn will correspond to the total numbers of each class in the ballot-box. It is reasonable to expect that when there are 10,000 ballot papers in an urn the composition of the first thousand drawn out will nearly be the same as that of any other thousand, or of the whole 10,000. The amount of this probability may be determined mathematically, and is very great.

This fact was clearly seen by Mr. Andrae, the statesman by whom the method of preferential voting was introduced into Denmark in 1855, and a mathematician of undisputed eminence. In answer to an objection of the kind now under discussion, he replied: "If this law of mine had already been in operation over the whole of Europe (including Turkey), for a period of 10,000 years, and if the elections in every part of Europe to which the law was applied were to take place, not every one, or three, or seven years, but every week in regular repetition, these elections throughout Europe, at the rate of a general European election per week, would still have to go on for more than a thousand times the period of years already stated; that is to say, for more than a thousand times ten thousand years, before the chances would be equal that the voting papers should come out of the urn in the order required to form the basis of this problem. Although, therefore, the supposed combination is, mathematically speaking, only an enormous improbability, yet, practically speaking, it is absolutely impossible."[2]

To state the matter more exactly, and as the result of an independent mathematical investigation, it appears that in the case we have stated, if 4000 voting papers were drawn out of A's heap at random, instead of the papers being carefully sorted and proportionately divided, the probability is that neither B nor C would gain or lose more than 11 votes. In other words, it is just even betting that the number of AB votes in the 4000 drawn would lie between 2411 and 2389 (inclusive), and consequently that the number of BC votes will lie between 1589 and 1611. The odds are more than 3 to 1 neither B nor C would gain or lose more than 20 votes, i.e. that the number of AB votes drawn will lie between 2420 and 2380; more than 10 to 1 that neither would gain or lose more than 30 votes; just 50 to 1 that neither would gain or lose more than 40 votes; and about 2000 to 1 that neither would gain or lose more than 60 votes. If the number of classes were larger or the number of votes to be drawn smaller, the effect would be much less. It will thus be seen that it is only in the case of very closely contested elections that the element of chance can affect the result. It will also be observed that the element of chance will not be of importance as between the different parties, but only as between different individual candidates of the same party, since in almost all cases the electors who are agreed upon the candidate they most desire will also put for their second choice candidates of the same party.

In closely contested elections it must, of course, be admitted that as a result of this method, chance might decide which of two candidates of the same party should be elected. But in closely contested elections in large constituencies so many elements of chance are always and necessarily involved, that the introduction of a fresh one does not, in reality, make the result more arbitrary. Putting aside all the slight influences which at the last moment decide a score or two of featherweight votes, and assuming that every voter is profoundly convinced of the truth of his opinions, there remains the question of boundaries. A slight change in the line of the boundaries of the constituency might easily make a difference of fifty votes—a larger difference than what we are concerned with. To carry the dividing lines from North to South instead of from East to West, would, in many localities, completely alter the character of the representation.

These are, in reality, matters of chance, and more arbitrary in their nature than the order in which voting papers are drawn from an urn.

(2) Method of Eliminating the Chance Element

If, however, special precautions are still thought necessary, the following method of counting the votes appears to reduce, as far as practicable, the element of chance involved in the transfer of superfluous votes:—

The whole set of voting papers of the constituency being mixed, the papers, not yet unfolded, are drawn out one by one. Each is stamped, as it is drawn, with a corresponding number, 1, 2, … in order. It is then unfolded, and sorted according to the names of the candidates marked first and second upon it. Suppose there are six candidates, A, B, C, X, Y, Z; the votes of any candidate, A, will be sorted into six heaps, viz., A votes (i.e. votes where A only is voted for), AB, AC, AX, AY, and AZ votes. If A is found to have received more votes than he requires, the order in which the votes will be counted to him will be as follows: Use first the A votes, then use up those heaps where the second name also is that of a candidate who has received more than the necessary minimum. If these heaps give A more than he requires, take the same proportion out of each of such heaps, taking out of each heap the last drawn votes first. If, however, these heaps are used up without giving A as many votes as he requires, take an equal proportion of the votes of each of the remaining heaps—taking out of each heap the last drawn votes first.

Example.—Take an election where 6000 is the necessary minimum, and suppose A has 8650 votes, composed as follows:

A 600
AB 2,700
AC 4,500
AX 50
AY 200
AZ 600
——-
8,650

Using first the 600 A votes, we are left with 5400 to make up out of the remaining heaps.

1. Suppose B and C have received the quota. The 5400 can be taken from their heaps exclusively, for in their two heaps are 7200 votes; the proportion to be taken from each heap is therefore 5400 out of 7200, which is three quarters. Thus we make up A's number thus:—

A votes 600
Three-quarters of 2,700 AB " 2,025
Three-quarters of 4,500 AC " 3,375
——-
6,000

And transfer the remainder (the AB and AC votes transferred being those stamped with the lowest numbers).

2. Suppose B and X have received the quota. Their two heaps amount to 2750 votes. Using these up, there remain 2650 votes to be made up out of the AC, AY, and AZ heaps. These three heaps together contain 5300 votes; and the proportion to be taken from each heap is 2650 out of 5300, or half. Thus A's number is made up as follows:—

A votes 600
AB " 2,700
AX " 50
Half of 4,500 AC " 2,250
Half of 200 AY " 100
Half of 600 AZ " 300
——-
6,000

And the remaining votes of each of the three last classes—being those stamped with the lowest numbers—will be transferred.

It will be observed that the element of chance is not wholly excluded, since the question, which papers out of the AC heap are transferred, is left to depend upon the order of drawing. To exclude chance wholly, these would have to be sorted into heaps according to the third name upon them, and an equal proportion taken from each heap. The figures in the first half of this paper are sufficient to show that such trouble would be wholly superfluous.

[Footnote 1: This Memorandum is published by permission of the Rt. Hon. J. Parker Smith. Although written in 1884, the arguments still apply. The method described in the second part of the paper has been adopted in the Municipal Representation Bill (see Appendix VII.), but the method of application differs in detail.]

[Footnote 2: Quoted by Mr. (afterwards Earl) Lytton in his Report on the Election of Representatives for the Rigsraad.—House of Commons papers, 1864, vol. 61, p. 24 of No. 7.]

APPENDIX VII

THE SINGLE TRANSFERABLE VOTE

SCHEDULE TO MUNICIPAL REPRESENTATION BILL, 1910

THE FIRST SCHEDULE[1]

RULES FOB THE TRANSFER OF VOTES AND FOR ASCERTAINING THE RESULT OF THE POLL

Arrangement of ballot papers.

1. After the ballot papers have been mixed, in accordance with the rules contained in the First Schedule to the Ballot Act, 1872, the returning officer shall draw out all ballot papers which he does not reject as invalid, and file in a separate parcel those on which the figure 1 is set opposite the name of the same candidate. The returning officer shall then count the number of papers in each parcel.

Ascertainment of quota.

2. The returning officer shall then add together the numbers of the papers in all the parcels and divide the total by a number exceeding by one the number of vacancies to be filled, and the result increased by one, disregarding any fractional remainder, shall be the number of votes sufficient to secure the return of a candidate, herein called the "quota."

Candidates with quota elected.

3. Any candidate whose parcel contains a number of papers equal to or greater than the quota shall be declared elected.

Transfer of surplus votes.] 4.—(1) If the number of candidates elected under the last rule shall not equal the number of vacancies, the returning officer shall as far as possible transfer from each elected candidate the votes (if any) in excess of the quota (herein called surplus votes) to the candidates indicated on the ballot papers as next in order of the voters' preference, excluding candidates already declared elected. The votes of the candidate having the largest number of votes shall first be dealt with, and the particular votes to be transferred shall be determined in accordance with the following regulations:—

(a) The returning officer shall arrange all the ballot papers in the parcel of the elected candidate on which votes capable of transfer are given by filing in a separate sub-parcel those on which a next preference is indicated for some one continuing candidate.

(b) The returning officer shall also make a separate sub-parcel of the ballot papers in the parcel on which the votes given are not capable of transfer.

(c) The returning officer shall count the ballot papers in each sub-parcel, and also the total of all the ballot papers containing votes capable of transfer.

(d) If the total number of votes capable of transfer is equal to or less than the surplus votes, the returning officer shall transfer all the votes capable of transfer.

(e) If the total number of votes capable of transfer is greater than the surplus votes, the returning officer shall transfer from each sub-parcel of votes capable of transfer the number of votes which bears the same proportion to the total of the sub-parcel as the number of surplus votes bears to the total of all the votes capable of transfer.

(f) The number of votes to be transferred from each sub-parcel under the preceding regulation shall be ascertained by multiplying the total of the sub-parcel by the number of surplus votes and dividing the result by the total number of votes capable of transfer. Fractional remainders shall be disregarded.

(g) The particular votes transferred from each sub-parcel shall be those last filed in the sub-parcel.

(2) The transfer of surplus votes shall be effected by making new sub-parcels of the ballot papers on which those votes are given, and adding those sub-parcels to the parcels (if any) of the candidates to whom the transfers are made, or, where any such candidate has as yet no parcel, a new parcel shall be formed for him from the papers transferred.

(3) All ballot papers in a parcel of an elected candidate not transferred under this rule shall be set aside as finally dealt with, and the votes given thereon shall thenceforth not be taken into account.

(4) If two or more parcels of elected candidates are equal in size, the returning officer shall decide which parcel he will first deal with under this rule.

(5) A transfer of votes under this rule shall not be made unless the surplus votes of the elected candidate, together with any other surplus votes not transferred, exceed the difference between the totals of the votes of the two continuing candidates lowest on the poll.

(6) This rule shall take effect subject to the provisions for filling the last vacancy herein-after contained, and if at any time it shall be possible to fill the last vacancy under those provisions, no further transfer under this rule shall be made.

Result of transfer.

5. After the transfer of the surplus votes of an elected candidate, any candidate who shall, as a result of the transfer, obtain the quota of votes, shall be declared elected.

Further transfer of surplus votes.

6.—(1) Unless and until the last vacancy shall have been filled under the provisions herein-after contained, if, after the transfers directed by Rule 4, there shall still remain a vacancy, and the votes of any elected candidate to whom a transfer has been made are in excess of the quota, the returning officer shall, as far as possible, take from the sub-parcel last transferred to that candidate a number of votes equal to the surplus.

(2) The particular votes to be taken shall be determined in accordance with the regulations given in Rule 4 hereof, in the same manner as if the votes included in the sub-parcel last transferred had been the only votes given to the candidate; the ballot papers so taken shall be added in separate sub-parcels to the parcels of the continuing candidates (if any) indicated thereon as next in order of the voters' preference, and the votes given thereon shall be transferred to those candidates accordingly. Where any such candidate has as yet no parcel, a new parcel shall be formed for him from the papers transferred.

(3) The remaining ballot papers in the parcel of the elected candidate (including the ballot papers taken from the parcel under Sub-Rule (1) on which the votes given are not capable of transfer) shall be set aside as finally dealt with, and the votes given thereon shall thenceforth not be taken into account.

(4) After any transfer of votes under this rule, any candidate who shall, as a result of the transfer, obtain the quota of votes shall be declared elected.

(5) The process directed by this rule shall be repeated until the last vacancy is filled, or until no candidate has any surplus votes, whichever shall first happen.

(6) If two or more parcels shall be equal in size, regard shall be had to the number of votes counted to each candidate under Rule 1, and the parcel of the candidate highest on that count shall first be dealt with, but if the numbers of votes on that count were equal, the returning officer shall decide which parcel he will first deal with under this rule.

(7) A transfer of votes under this rule shall not be made unless the surplus votes of the elected candidate, together with any other surplus votes not transferred, exceed the difference between the totals of the votes of the two continuing candidates lowest on the poll.

Distribution of votes of lowest candidate.

7.—(1) Unless and until the last vacancy shall have been filled under the provisions herein-after contained, if, after the transfers under the preceding rules, there shall still remain one or more vacancies, or, if no candidate shall have been declared elected under Rule 3, the returning officer shall exclude from the poll the candidate having the lowest number of votes, and shall distribute the votes capable of transfer on the ballot papers in his parcel among the continuing candidates next in order of the voters' preference. Any ballot papers in the parcel, on which votes not capable of transfer are given, shall be set aside as finally dealt with, and the votes given thereon shall thenceforth not be taken into account.

(2) If in any case the total of the votes of the two or more candidates lowest on the poll together with any surplus votes not transferred is less than the votes of the next highest candidate, the returning officer may in one operation exclude those candidates from the poll and distribute their votes in accordance with the foregoing provisions.

(3) After the distribution under this rule of votes capable of transfer, any candidate who has received the quota shall be declared elected.

(4) The surplus votes of any candidate elected under this rule who has received more than the quota shall be distributed in the manner directed by and subject to the conditions of the last preceding rule.

Further distributions.

8. The process directed by the last rule shall be repeated on the successive exclusions one after another of the candidates with the lowest numbers of votes until the last vacancy is filled either by the election of a candidate with the quota or under the next following rule.

Filling the last vacancy.

9.—(1) When the number of continuing candidates is reduced to the number of vacancies remaining unfilled, the continuing candidates shall be declared elected.

(2) When only one vacancy remains unfilled and the votes of some one continuing candidate exceed the total of all the votes of the other continuing candidates together with any surplus votes not transferred, that candidate shall be declared elected.

(3) When more than one vacancy remains unfilled and the votes of the candidate, who, if all the vacancies were filled by the successive elections of the continuing candidates with the largest numbers of votes, would be the last to be elected, exceed the total of all the votes of the continuing candidates with fewer votes than himself together with any surplus votes not transferred, that candidate and all the other continuing candidates who have not less votes than himself shall be declared elected.

(4) When only one vacancy remains unfilled and there are only two continuing candidates, and those two candidates have each the same number of votes and no surplus votes remain capable of transfer, one candidate shall be declared excluded under the next following rule and the other declared elected.

Provisions for exclusion of candidates in special cases.

10. If at any time when a candidate has to be excluded under these rules two or more candidates have each the same number of votes, regard shall be had to the number of votes counted to each candidate under Rule 1, and the candidate lowest on that count shall be excluded, but, if the numbers of votes on that count were equal, the returning officer shall decide which candidate shall be excluded.

Public notice of transfers.

11. The returning officer shall record and give public notice of any transfer of votes made under these rules and of the total number of votes counted to each candidate after any such transfer in addition to the particulars prescribed by Rule 45 to the First Schedule to the Ballot Act, 1872. Such public notice may be in accordance with the form given in the appendix to these rules.

Recounts.

12.—(1) Any candidate or his agent may at any time during the counting of the votes, either before the commencement or after the completion of the transfer of the votes (whether surplus or otherwise) of any candidate, request the returning officer to recount the papers then comprised in the parcels of all or any candidates (not being papers set aside as finally dealt with) and the returning officer shall forthwith recount the same accordingly. The returning officer may also at his discretion recount votes either once or more often in any case in which he is not satisfied as to the accuracy of any previous count. Provided that nothing herein shall make it obligatory on the returning officer to recount the same votes more than once.

(2) If upon an election petition—

(i) any ballot papers counted by the returning officer are rejected as invalid,

or

(ii) any ballot papers rejected by the returning officer are declared valid,

the court may direct the whole or any part of the ballot papers to be recounted and the result of the election ascertained in accordance with these rules.

(3) Except as in this rule expressly provided, no recount shall be had whether on an election petition or otherwise.

_Determination of questions as to transfers.

13.—(1) If any question shall arise in relation to any transfer, the decision of the returning officer, whether expressed or implied by his acts, shall be final unless an objection is made by any candidate or his agent before the declaration of the poll, and in that event the decision of the returning officer may be reversed upon an election petition.

(2) If any decision of the returning officer is so reversed, the transfer in question and all operations subsequent thereto shall be void, and the court shall direct what transfer is to be made in place thereof, and shall cause the subsequent operations to be carried out and the result of the election to be ascertained in accordance with these rules.

Definitions.

14. In these rules—

(1) The expression "votes capable of transfer" means votes given on ballot papers on which a further preference is indicated for a continuing candidate. Provided that a vote shall be deemed not capable of transfer in any case in which—

(a) The names of two or more candidates (whether already excluded from the poll or declared elected or not) are marked with the same figure and are next in order of preference, or

(b) The name of the candidate to whom the transfer is to be made or of some candidate (whether continuing or not) higher in the order of the voters' preference is marked

(i) by a figure not following consecutively after some other figure on the ballot paper, or

(ii) by two or more figures.

(2) The expression "continuing candidates" means candidates not already declared elected or excluded from the poll.