H.—4l

4

But the following proposal would have no such effect as that last mentioned, nor would it involve any departure from principle, nor can I see any objection to it in practice. It is this : that all papers with the transfer value 1 should be dealt with as if they belonged to one bundle ; similarly, all other papers with the same transfer value (less than 1) should be transferred at one step. This would shorten the counting, although of course not so much as the first suggestion. It would present in some cases an apparent anomaly—namely, that the total number of votes for the several transferees might be more than the number of votes given up by tho candidate who was being excluded ; the excess would be really a recovery of votes lost by fractions, the total for which could be reduced accordingly. Time taken in the Counts. Seats. Candidates. Time taken. Christchurch City Council .. .. .. 16 30 33 hours. Harbour Board .. .. .. .. .. 4 9 8 ~ Hospital Board .. .. .. .. ..7 11 13 „ Number of papers counted in each case, about 17,000. The time taken in the count varies, of course, with the number of papers (or voters), number of seats, and number of candidates. The time required for the first count varies as the total number of voting-papers. The nearest indication of the time required for the remaining counts seems to be the excess of candidates over vacancies—in other words, the number of candidates to be excluded : to give a still better index, this number might be multiplied by the number of candidates elected before the first candidate was excluded. Time taken in the Scrutiny op Votes. It may be of interest to note that the number of separate examinations of papers in the sorting and checking of the 174 effective counts of the City Council election was over 65,000 (besides about 1,000 examinations of papers in the non-effective counts referred to above). The following table affords a comparison of the work involved, and the time taken in doing it :■ —

Table A.

The expense of the election of course depends partly on the length of time taken over the counts, if the staff is the same. In Tasmania, apparently, a much larger staff was employed. In elections conducted according to systems at present in vogue the public has been led to expect an early publication of the result of the poll; accordingly, a new system of voting stands in danger of being condemned if the counting of the votes takes longer than the older system to which we have grown accustomed ; for the moment, the fact that the new system gives a fairer representation of the mind of the constituency is apt to be lost sight of. This prejudice will no doubt disappear upon mature consideration, for there can be very few elections in which delay for a few hours in announcing the results can be of material importance in comparison with the benefits gamed —if these are conceded. This is not the place to give the arguments in favour of proportional representation ; they may be found in any text-book upon the subject. As regards the expense of the election, inasmuch as the counting of the votes is only a part, and not the largest part, of the process, it is probable that the additional expense involved would not be great. The system does indeed appear to become more complicated when the number of seats and the number of candidates is large ; this is not true as far as it concerns the counting, for the longer time taken in the count is due not to any new methods or processes required, but to the continued repetition of the comparatively simple processes involved in transferring surplus votes from elected candidates, and in transferring to the next preferences the votes of candidates excluded on the ground that they are the lowest and have therefore no chance of being elected. To the voter it is probably more difficult to express intelligently the order of his preferences when there are, say, thirty candidates than when there are, say, only ten or fifteen. On this ground, therefore, much may be said for constituencies electing not more than perhaps six or seven members. Even this contention may be modified by reference to tables C and D below, which show that the later preferences exercise very little influence upon the election; so that a voter who puts down the first six preferences and no more is omitting nothing or almost nothing that would really affect the result, unless indeed the last candidates elected were very near to the highest of the rejected candidates. On the ground of expense, it is almost certain that in Christchurch it would have cost more to hold the election in two constituencies electing eight members each than it did to hold one election for the whole city electing sixteen members.

Number of Seats. Number of Candidates. Number of Valid Votes. Number of Examinations of Papers. »-»| Sat Christchurch City Council Harbour Board Hospital Board L6 4 7 so 9 11 16,677 15,488 15,507 Over 65,000 57,000 55,000 174 29 28 33 8 13 Tasmanian Elections, 1909. Darwin Bass Denison 6 6 6 9 L3 L6 9,405 9,070 11,337 % Over 19,000 16 54 109 II 15 11