in almost all cases a sure prognostication of what is to happen, as is the result in Maine shortly afterward; and it is not a superstition. On the contrary, it is founded upon a philosophical principle that cannot be successfully disputed. Mr. Bryan was as surely defeated in 1896, when Vermont gave Grout thirty-eight thousand majority, as he was when the polls closed in November. In order to maintain this proposition it is not necessary to suppose that a single voter anywhere in the country changed his political intention as a consequence of the Vermont election, or that any man, previously undecided, determined to 'jump on the band-wagon.' The real reason is that men in Indiana, in Idaho, and in Vermont, influenced by the same events, actuated by the same motives, and listening to the same arguments, act in the same way. Some of them, of course, are drawn in one direction, others in the opposite direction, according to what manner of men they are, and what original opinions and tendencies they represent.

Grant that Vermont is not, politically speaking, a typical American community, yet it does contain all sorts and conditions of men, although in different proportions from the distribution in many other communities. When, therefore, it appears that there has or has not been a perceptible political change, caused by a movement by one or more of the many classes of population from one party to the other, the country is supplied with a reasonably trustworthy view of the state of political sentiment in Indiana, Idaho, and elsewhere. Events, it is true, may occur between September and November that will affect and modify political action all over the country, and in Vermont as well; but they must be events, and not merely transitory waves of sentiment.

We frequently see in the newspapers, a few weeks before the election, statements by political correspondents that the prospects of this party or the other have improved or grown less promising during the week past, or that there is now a perceptible drift toward this candidate or that. Do readers ever stop to consider what this means, or whether there can possibly be any foundation for such statements? Does any one suppose that there is ever a considerable body of voters in any State who are undecided how they will vote, and who secede in a flock from their party one week, and return to it the next? Or if there were such a body, can any one suggest how the sapient correspondents ascertain the fact? It may not be an unjustifiable conjecture that the sole basis of such statements as we are considering is the state of mind, optimistic or the reverse, of the committee chairman or the local politician who communicates information as to the political situation to the newspaper interviewer. The chairman may have received a despondent letter from a county manager, and from it may conclude that the cause is in a bad way in that part of the State. Or he may have had a good night's rest and an excellent breakfast. His mood will determine the character of his outgivings. But, in reality, nothing has happened; or if it has, he does not know it.

It may be asked, if this be sound political reasoning, why the frantic campaigning and stump-speaking of the September and October preceding the election? If the race has been decided, why does one party not rest on its oars and the other give up and row back to the stake-boat? There is need that some old hand on the stump, who is also a good observer, should present to the country an analytical and philosophical study of the purpose and the result of campaign oratory. To

the superficial outside observer, what should be, and ostensibly is, its main purpose, the conversion of political opponents, is seldom accomplished, even to a limited extent. How could one expect it to be? Unless the speaker is a man of great power and reputation, the audiences he attracts consist almost exclusively of voters who are already enlisted in his party and do not need to be convinced or converted. On the other hand, if he is a person of national prominence or noted for his eloquence, he has some, perhaps many, political opponents among his hearers. But they do not go to his meetings with open minds, but out of curiosity; and the views, principles, and intentions which they take to the meeting they carry away unchanged.

The most successful stumping tours in our political history, so far as the number addressed was concerned, and the most spectacular, were those of Mr. Blaine in 1884, and those of Mr. Bryan in his three campaigns. But the election returns at the close of the canvasses cannot be tortured, with the utmost mathematical ingenuity, into proving that by their eloquence an appreciable inroad was made in the ranks of their opponents. Moreover, if personal observation goes for anything, one might appeal to the common experience of every man with the question: Did you ever meet or know of a voter who was converted from one party to another by a stump speech?

Undoubtedly 'spell-binding' has its uses. If not, campaign committees would have found it out long ago and abandoned the practice, instead of organizing political meetings in every hamlet and providing as speakers a few stars and a multitude of third-rate men. The manufactured enthusiasm of those who attend the meetings probably has an influence in dissuading doubting and hesitating voters from

deserting their party. It also certainly has the effect of bringing indifferent citizens to the polls on election day. It may be that experienced campaigners have been able to discover some other benefit, direct or indirect, of the system; but those just mentioned are the only ones that are obvious to the political student who is not in the inner circle of management.

The party that is at any time in the minority, and out of power, hopes for and predicts a 'landslide.' Now there is one test, heretofore infallible, to be applied to political opinion at any given time. A landslide, or a fairly stable condition of the political sentiment of the country, can be foretold with even more confidence than an inspection of the barometer gives us in respect of the weather. A political upheaval to put it in paradoxical form - does not originate from below, but from above. It would be difficult to cite an important overturn in national politics which was not foreshadowed by an open revolt of party leaders, and led and managed by them. Small variations in close districts and states do take place without the preliminary symptom just mentioned; but we are speaking now of changes that may be described as revolutionary. The fact might be illustrated by numerous examples. Indeed, as is implied by the form of the statement above, every overturn furnishes an example. But it will be sufficient to mention a few of them.

The revolt against Jacksonism which resulted in the election of Harrison, in 1840, was forecast by the secession of such Democratic leaders as Tyler, and Hugh White, and Berrien, and Mangum. Cass was defeated, in 1848, by the defection of Van Buren and many other leaders. The election of Lincoln was preceded by a wholesale desertion to the new Republican party

of a large group of senators and other prominent men. The movement which resulted in the defeat of Blaine was originated and engineered by life-long Republicans. The campaign of 1896 occurred but yesterday. It was characterized by two 'landslides,' one in the West led by Teller and other senators; the other in the East, where a host of leading Democrats set the example of revolt from the free-silver movement. Prior to the election of 1908 the Democrats predicted a landslide here, there, and everywhere. But there were no prominent men of the other party who were moved by principle to desert to the other side, none who scented a revolution which promised profit to those who should take part in it; and there was no landslide anywhere.

All these desultory and disconnected remarks refer to the period before the election. One or two important matters that arise out of the situation when the votes have been cast, remain to be considered.

On many occasions, after a presidential election had been held and the returns were in, curious or alarmist statisticians have put forth calculations showing that the change of a small number of votes in one state, or two or three states, would have given victory to the defeated candidates. If 2554 men in New York who voted for Polk, in 1844, had voted for Clay, Clay would have been elected. Or the same result might have been reached if 3167 Pennsylvania Democrats had shifted to Clay, and if there had been no Plaquemines Fraud. The case of Blaine, in 1884, is hardly in point, because, although a shift in New York of 575 votes as they were counted would have elected him, there is a strong probability at least that he did actually have a plurality of the votes honestly cast in that State. But in 1888, although Cleveland had a popu

lar plurality of almost 100,000 he had only 168 electoral votes, whereas Harrison had 233. The vote of New York was: for Harrison, 650,338; for Cleveland, 635,965. Plurality for Harrison, 14,373. So, and this illustrates the method under consideration, if 7187 of the Harrison votes had been cast for Cleveland he would have had the thirty-six electoral votes of New York, which would have made his total 204, and left only 197 for Harrison.

That is all true; but there is included in all such calculations an assumption that such a change can take place in one state without being reflected by a corresponding change elsewhere. That is contrary to the principle that similar persons, acted upon by the same influences, act in the same way. In the case just cited it is proposed to consider the consequence of a bolt from the party candidates by more than one in a hundred of the Republican voters. In that case we should anticipate and should find a bolt of about one per cent of all the Republican voters in the country, and the net change in that case would have been not seven thousand, but many times that number, and Cleveland's plurality would have been more than doubled. The loser of a hand at whist sometimes tells what he would have done if he had only had another trump. But that change in his own. hand would have altered all the hands.

Inasmuch as it would have required a transfer to Bryan of more than seventy-seven thousand Republican votes, carefully distributed in eight states, to reverse the result of the last election, we did not hear the old story that the minority party came near to success. But the statisticians have indulged themselves in a consideration

one can hardly call the comments of most of them a study - which it may be worth while to examine, although any subject which, like this, involves

an arithmetical analysis of figures, is necessarily dry.

The point that is made by them is that the total vote in 1904 showed a remarkable decrease, as compared with that in 1900, and that the increase in 1908 over 1904 was by no means as large as the apparent increase of population would lead one to expect. The facts are accurately stated, but the suggestion that they are not capable of easy and simple explanation is not justified. The total vote of the country at the last three elections was as follows:1

1900

tions are slightly different, - North Carolina and Tennessee, the result is so well-assured in advance that whatever political effort is made locallyfor the national committees take no part in it is needless on the part of the Democrats and futile on the part of the Republicans. We may say, then, that whether a light vote, or one comparatively lighter, is cast in these eleven states is purely a matter of accident, and wholly without significance. The total vote in the eleven Southern states at the last three elections was as follows:

1904

1908

The decrease in 1904 as compared with 1900 was 3.27 per cent; the increase in 1908 over 1904 was 10.16 per cent; and the increase in the eight years from 1900 to 1908 was 6.56 per cent. If then we do not go beyond these figures the point mentioned above is proved, for the increase in population during the eight years has undoubtedly been more than seven per cent. But it will not do to rest upon such a general statement, for that is to disregard wholly the remarkable aloofness of the Southern states from the party contests of the rest of the country. There are nine such states in which there is never the semblance of a canvass. Not to burden this article with too many figures it may be said that the largest vote given in these states at any one of the last three elections, that of 1900, represented but 37.3 per cent of the males of voting age, and only 60.4 per cent of the white males. There is absolutely no inducement for Democrats to go to the polls, and—if that were possible less than none for the few Republicans who may be allowed to vote. In two other states where the condi

1 These are the figures of the New York Tribune Almanac.

Comparing these figures with those for the whole country, we see that the decrease during the first four years was just above half a million, which was rather more than that in the country as a whole; and that the increase in the second period, 200,000, compares with 1,362,000, in the whole country.

There are five 'border states,' Delaware, Maryland, West Virginia, Kentucky, and Missouri, where the contest is as strenuous as it is anywhere in the United States. Here are their total votes:

A slight increase of a little more than one per cent in the first period, followed by an increase of a little more than seven per cent in the second period, and an increase for the eight years of 8.54 per cent, which is quite as large as the increase of the voting population, if we bear in mind the fact that a large part of the increase of the total population in recent years has been made by immigrants who do not always come to stay, and who do not always become citizens if they do stay.

That

Statistical calculations of this sort are necessarily dry; but those who have followed the foregoing analysis will perceive that little is left of the point which we set out to examine. little is the fact that in the Northern States the total vote did not increase in 1904, as compared with 1900, so much as the natural rate of increase of the voting population would lead one to expect. But the fact involves no mystery for those who observed and remember the characteristics of the last three presidential canvasses. Although the statement involves what every one knows, or ought to know, it may be put briefly and broadly.

The canvass of 1896 was characterized, as has been already remarked, by two distinct movements: Republicans by the thousand going over to the Democrats, Democrats revolting against the party platform and candidates. Almost all the Northern States west of the Missouri River gave their electoral votes to Bryan; every Northern State east of that river voted for McKinley, generally by very large majorities. In 1900 the situation was more nearly normal. There was a great decrease of the Bryan vote in the Far West, a considerable increase in the Eastern States; but the vote for Mr. Bryan was still in a marked degree a vote of radicals, who had full control

VOL. 110 - NO. 4

of the party and dictated candidates and policies.

This brings us to the canvass of 1904, and to the explanation of the comparatively light vote of that year. A variety of influences affected the result. There was, first, the exceeding popularity of Mr. Roosevelt; secondly, the voluntary or enforced effacement of the radical element of the Democratic party; thirdly, the absence of any paramount' issue. They all tended in one direction. They produced an enormous increase of the Republican vote Republican vote-more than 400,000. A vast number of radical Democrats manifested their displeasure at the change in the tone of their party, by either voting for Mr. Roosevelt or neglecting to vote at all, and the returns showed a loss of more than a million and a quarter Democratic

votes.

It is, of course, impossible to estimate the extent of the defection, or to guess how many 'bolted' the ticket, and how many failed to vote. But we see the resultant of all the forces, and it is precisely that which coincides with the observation of every man whose eyes and ears were open in 1904. The canvass of 1908 saw the radicals again in control of the Democratic party, and it saw also a much more kindly and tolerant spirit toward Mr. Bryan on the part of conservative members of the party. Moreover, there were local contests in such states as New York, Ohio, Indiana, Illinois, and others over the governorship, with an advantage in every case on the Democratic side. This led to a spirited contest, an enlarged vote, and a sympathetic increase of the strength of the Democratic electoral ticket. General result: a slightly larger Republican total than ever before, caused by an increase of moderate amount in the Far West and a decrease in some states of the East and