were elected by Republican as well as by Democratic voters, but under the Richards law are in that case responsible only to the Democrats. Similar situations obviously may occur in other localities and in the state at large, and it may therefore seem that the law lays down a principle of doubtful value. The theory of the Richards plan, however, is that the real selection after all is not by the voters in the general election, but by each party group in making its nominations, and that each party ought properly be held responsible, therefore, for those officials whom it has placed before the general body of voters.

So far as the writer is aware, no recall proceedings have yet been instituted under the Richards law. The application of the recall provision has been tested, however, in the case of a county judge (Judge Burns of Deuel County), who, in an attempt to avoid ouster proceedings brought under another statute, claimed that he could be removed only through the party recall. The Supreme Court of the state, in denying his plea for a writ of prohibition, held that "the recall procedure (of the Richards law) does not purport to cover the entire field of the matter of removal from office and is therefore only cumulative to, and not inconsistent nor in conflict with, the general statutes upon removal." 27 Although the court did not pass directly upon the constitutionality of the party recall, the inference from its decision is that the provision is valid.

EXPENSE

One of the serious criticisms that has been made against the Richards primary is that it is unduly expensive. It is quite obvious that the voluminous machinery of initial and primary elections must be a source of considerable

expense to the taxpayers, but probably

27 See Sioux Falls Press, Jan. 27, Mar. 3, 1922.

not much greater than any other legally regulated primary system. In addition, however, the novel features of the Richards primary are all a source of expense. The proposalmen and the party committeemen are paid a mileage of five cents each way for attendance upon all necessary meetings, which in the case of the proposalmen in 1921, amounted to an average of $18 apiece, or a total for the three proposal conventions, of about $10,000. The candidates required to debate were allowed a mileage of ten cents for all necessary travel in that connection. Presumably General Wood collected mileage for travel between Chicago (then his military headquarters) and Pierre, Senator Poindexter between Washington and Pierre, Mr. Gerard between New York and Sioux Falls, and Mr. Monroe between Maywood, Illinois, and Sioux Falls. In addition, mileage must have been paid for all the travel required by the 28 gubernatorial debates, altogether amounting to a tidy sum.

The heaviest expense, next to the proposal conventions and the primary election, was that necessitated by the publicity pamphlet. Candidates who desired the insertion of cuts and biographies were required to contribute $100 for both or $50 for either, but otherwise the expense of publication was borne by the state. This amounted in 1919-1920 to $4,500. In addition, the expense of mailing was borne by each county, which in Minnehaha County amounted to $150, exclusive of clerical and other expense, and might be estimated at about $1,500 for the entire state.

The total expense to the state in 1919-1920 for mileage, publicity, and incidentals can be estimated at about $20,000,28 and in addition an expense

supplied by the Secretary of State, in a letter of

28 These figures are estimated from information

Dec. 8, 1921.

casion but one sustained the Richards law; but the struggle has confirmed the constitutional right of the legislature (at least in South Dakota) to repeal or amend an initiated law at will. Although the legislature of 1921 accordingly did repeal several of the novel features-the provisions for a publicity pamphlet, public joint debates, party endorsement for appointive positions, and the postmaster primary-the salient provisions of the Richards plan with respect to the proposal and selection of candidates and issues remain intact, and are likely to be respected. The machinery is cumbersome and complicated, but has worked rather contrary. On the whole, the experiwell in spite of dire predictions to the ment has been unusually interesting and, as the writer feels, well worth while. Its continued operation will be watched with keen interest by students of government and party politics.

T has been repeatedly asserted that

in the primary. This, of itself, might not be an unanswerable argument against the primary, but certainly the apparent apathy of the voters does not aid the cause of the adherent of the direct primary. Newspaper comments following a primary election in Indiana bear witness that the average editor is inclined to condemn the primary on this ground. "Light Vote Cast," or "Only 50 Per cent Vote," and similar headlines have given the general impression that there is no interest in the primary.

Such, however, is not the case. A statistical analysis of Indiana primary returns, from vote for state officers down to the vote in townships and precincts, reveals an astonishing interest in the primary wherever the candidates nominated at such primaries have a chance of success in the election.

INTERPRETING THE PRIMARY VOTE

There seem to be several misconceptions as to the conclusions which can be drawn from what appears to be a light vote in the primary. In the first place, the Indiana primary vote over the state averages from 50 per cent to 54 per cent of the vote at the election. This is, in fact, a relatively high primary vote. The difficulty is that the primary is compared with the election. It should be compared rather to the old primary system. So far as can be ascertained, a 10 per cent vote was considered a good vote under the old system, and there are

many tales in Indiana of primaries in which a mere handful of voters named the slate. A 50 per cent vote today then is actually an increase of 400 per cent in interest in the primary and in the extent to which the average voter is participating in party affairs. In many counties in Indiana one party or the other casts a primary vote of 70 per cent, or an increase of 600 per cent in interest and participation.

Moreover, in comparing the primary with the election, no allowance is made for the independent vote. While the independent may participate in Indiana, he is not a party man and should not be expected to participate in an election within the party, and his failure to vote in the primary in no wise enters into the merits of the direct primary. If we accept Mr. Merriam's recent estimate on independent voters, we should not expect much over a 75 per cent vote in the primary. In Indiana the proportion of independents is probably lower, and an 85 per cent vote might be expected. On this basis the present 50 per cent vote in Indiana is nearly 60 per cent of the vote that might be possible. And a 60 per cent voluntary vote, with no party machinery to drag the voter to the polls, is, after all, a good-sized vote.

POLITICAL GEOGRAPHY MUST BE

CONSIDERED

But an analysis of Indiana counties reveals even more interest and participation in the direct primary. It is useless to add up the total vote in the state and draw conclusions from that. The political geography of the state

There are in Indiana 37 counties which are strongly Republican and 32 which are strongly Democratic. That is, a study of the 19 general elections in Indiana in the past 38 years demonstrates that all of these 69 counties have been carried by the dominant party in at least 14 out of the 19 elections. And in many cases the majority or plurality of the dominant party has been so great that it required the Progressive split of 1912 or the landslide of 1920 to turn the plurality into the opposite column. In these 19 elections, eleven counties have always gone Democratic; three have always been Republican. Five others were Progressive in 1912, but have otherwise been Republican. Six others went Democratic in 1912 only.

Such strongholds of party strength must be considered in any analysis of the primary, and the conclusions which can be drawn from such counties seem to be of great importance. In many of these counties the nomination is always equivalent to an election. They can be compared to the states

cratic primary is far more important than the election. And these 69 counties constitute 75 per cent of the 92 counties of Indiana.

REAL INTEREST IN MANY COUNTIES

To illustrate the results of the statistical survey, two tables are given. The first shows ten of the strongest Democratic counties, giving the percentage of vote cast in the primary as compared with the vote at the election the same year by parties for 1916, 1920, and 1922. In 1920 the women voted in the election but did not vote in the primary. To make the percentages comparable with the others they have been consequently multiplied by two. This is not accurate, but does roughly account for the women's vote, and seems preferable to setting forth the figures without making some allowance therefor.

The table shows first that the Republican vote in these counties is usually small, and uniformly much smaller relatively than the Democratic vote. On the other hand, the Democratic vote is frequently surprisingly

high, averaging 77.2 per cent in 1916 and 70.3 per cent in 1920, in many cases rising above 80 per cent.

The fact is, of course, that there is no reason to expect a high Republican vote in these counties where for 38 years the Republicans have never won an election, and where their only influence can be in assisting Republican strength in counties joined to them for joint election districts.

However, in these counties, for the Democrats the real contest comes in the primary and the voters respond.

party is distinctly dominant, bears out the general conclusions of these two tables. There is a very real interest in the direct primary in 75 per cent of the counties in Indiana, but then usually in the primary of the dominant party only. In the other counties, those which may be called "doubtful," the percentage of vote varies somewhat according to the strength of the majority by which the more successful party has carried the elections; but in all cases of counties which can be won or lost by small pluralities,

The second table shows the same facts for ten strong Republican counties, and the conclusions from the first table are reaffirmed. The dominant party invariably polls a high vote in the primary-a very high votewhile the party which has no chance to win the election is not greatly interested, naturally enough, in nominating men whose defeat is certain. In fact, the vote of the minority party in these counties is far lower than for the minority party in the first table.

A similar analysis of all of the 69 counties in which one or the other

the primary vote has usually been high. From this analysis, emphasis on the 50 per cent vote cast over the state is obviously improper, as it includes the votes of one party in 69 counties where no one would expect anything but a low vote.

The inevitable conclusion is that the direct primary should be retained in Indiana because in 75 per cent of the counties it is usually for county offices more important than the election itself, and is the only opportunity for the voter to cast a ballot where it will be of determinative value.