unconstitutional.

Nor is it sufficient that a statute violates

This spirit is

the spirit supposed to pervade the constitution. too evanescent to be dealt with by the courts. Attempts have frequently been made to have acts declared unconstitutional because contrary to the fundamental principles of republican government. But this reason involves a purely political question, falling within the province of the legislature. These fundamental principles are vague and are subject to variation with changes of public policy. The best established general principles are subject to reasonable exceptions; and the courts are not justified in saying that a certain act is not such an exception.2

It is not necessary, however, to the unconstitutionality of an act that it be in conflict with some express words of the written constitution. It is equally void if in conflict with implied provisions, or if inconsistent with some provision which confers the power sought to be exercised by the legislature upon some other department of the government, or if there is a failure to observe the forms in accordance with which the authority is to be exercised. Where judicial power is granted to the judiciary, it is impliedly denied to the executive and legislature. The general assignment of governmental powers to the three departments is equivalent to the exclusive grant to each of the whole authority naturally pertaining to its character.

CHARLES B. ELLIOTT.

1 Sharpless vs. Mayor, etc., 21 Pa. St. 160-4, Block, C. J. "The rule of law upon the subject," says Judge Cooley (Constitutional Limitations, 204), “appears to be that except where the constitution has imposed limits upon the legislative power, it must be considered as practically absolute, whether it operates according to natural justice or not in any particular case.”

2 In the License Tax Cases, 5 Wall. 469, Chief Justice Chase said: "There are undoubted fundamental principles of morality and justice, which no legislature is at liberty to disregard, but it is equally undoubted that no court except in the clearest cases, can possibly impute the disregard of these principles to the legislature. This court can know nothing of public policy except from the constitution and the laws and the course of administration and decision."

THE

ON CENSUS METHODS.

HE present enumeration of the people of the United States, which will be known as the eleventh census, is the greatest piece of statistical work undertaken in modern times. It will cover sixty-five million people, scattered over more than three million square miles of territory; while the variety of information sought will be more exhaustive, and the proposed analysis of results more elaborate, than anything that has ever been attempted in America or Europe. A census is in itself a matter of profound interest. It is most commonly thought of, perhaps, as a sort of national "taking of stock," to show the progress of the community in population, wealth and well-being. From a scientific point of view, however, it is much more than that. It is an opportunity for sociological observation, which comes only once in ten years and which must be utilized punctually and to its fullest extent; for that particular opportunity will never return. It is like a transit of Venus or a total eclipse of the sun, for the observation of which astronomers make preparations months beforehand and travel thousands of miles. Moreover, each successive census makes one of a series, and if it is defective or distorted, it will, for all time to come, be a disturbing element in comparisons of social growth and development.1

The importance of these considerations is often lost sight of in this country, on account of the lack of scientific interest in statistical work and of scientific criticism of the results. In England a somewhat more favorable condition of things seems to

1 How such an opportunity for sociological observation may be lost, and lost forever, is seen in the failure to take a census in the state of New York in 1885, owing to the quarrel between the governor and the legislature. How such an opportunity may be imperfectly utilized is seen in the defeat of the proposed census law of 1870, by which the ninth census would have been taken according to the improved methods adopted ten years later,—the defeat being due to the personal feeling between Senators Sumner and Conkling. See The Nation, vol. x, p. 116.

259

exist. The Royal Statistical Society of London takes an active interest in all such work and numbers among its members many government officials.1 On the continent, the numerous statistical journals are constantly discussing statistical methods and the value of the official returns. The old statistical congresses, which formerly met every few years, attempted to bring about a uniform system of classification in different countries; and the new International Statistical Institute has undertaken the same work. At the head of the statistical bureaus in Europe are placed men of the highest scientific attainments, such as Farr, Ogle and Giffen in England; Engel, Von Mayr, Becker, Böhmert, Körösi and Keleti in Germany and Austria; Bodio in Italy, Bertillon in France, and many others. None but a scientific man could stand the stream of criticism that assails the work; while not only does the community get the best service, but science is also advanced.

In this country no such interest is felt. No economic association or statistical society petitions Congress in regard to the scientific scope of the census.2 Additions are made to the schedules by Congress, but they are either for political effect or are due to momentary interest in a particular question, and they are often forced upon the superintendent without regard to their permanent usefulness or to the danger of over-burdening

1 The society "has since the year 1840 appointed a committee before every census to inquire into the various questions and report to the council. In consequence of the reports of these committees, the society has from time to time made important representations to the government of the day, which, while they have disappointed many by not effecting all that the society had hoped to achieve, yet have resulted in very important improvements. The most notable was in 1840, when the action of the society led the government to withdraw its bill and introduce another, modelled on the lines suggested by the council." Journal of the Royal Statistical Society, vol. 52, p. 440; see also vol. 51, p. 816.

2 It is a curious fact that at the beginning of the century very considerable interest seems to have been felt in the scientific possibilities of the census. Just before the enumeration of 1800, the American Philosophical Society, of which Thomas Jefferson was president, memorialized Congress recommending a more extended age classification, distinction between natives, foreign-born and aliens, distinction of occupations and professions, etc., the society believing the "duration of life to be longer here and the increase of population more rapid than elsewhere." The Connecticut Academy of Arts and Sciences, of which Timothy Dwight was president, petitioned for the same things in order to "collect materials for a complete view of the natural history of man

the enumerators.1 No adequate criticism follows the census. Disappointment is often felt at local results, and charges of partiality or even of corruption are freely made, because the returns do not show the population or the industrial activity which the imagination of a particular community had fondly pictured.2 The census office immediately assumes an attitude of defence, because its integrity is assailed. It learns nothing, because the criticism is on results, not on methods; that is, on its honesty, not on its skill. No new light is thrown on the subject. The superintendent of the next census, if he is of a combative disposition, simply presses forward on the old lines, prepared to endure the abuse which will follow; or, he unconsciously relaxes the severity of his methods, in order to get results that will "satisfy the people."

This sort of criticism, besides being grossly unjust to honest public officials, is in two respects misleading and injurious. In the first place, it entirely ignores the great question of the proper scope of a census. The most important result of experience in census-taking is the knowledge as to what a census can and ought to do, and what it cannot do and ought not to attempt. A distinguished French statistician has enunciated certain dicta on this subject which are well worthy of consideration, both by census officers and by the general public — including our law makers.3 There are certain facts which, owing to the nature of things or to the particular disposition of a people, cannot be ascertained. Questions touching personal religious belief or opinion, as well as questions in regard to personal indebtedness or private income, are apt to be resented as inquisitorial. Further, there are certain facts which can be ascertained but

and society in this country." Congress paid no attention to either petition. Garfield Committee Report, Second Session, 41st Congress, House Reports, vol. i, no. 3, p. 35. 1 Scientific criticism would probably reduce rather than increase the number of inquiries. The schedule is much too large and demands detailed answers which are a burden to both the enumerator and the enumerated. See Walker, article Census, Encyclopædia Britannica. Yet every superintendent has to fight to keep additional inquiries off the schedule, especially such as would awaken suspicion or antagonism and injure the whole return.

2 See Tenth Census, vol. ii, Introduction, p. xl.

3 Levasseur, La Population Française, reviewed infra.

which are not worth knowing; as, for example, the personal habits of individuals. Finally, there is a great list of things which can be known and which would be interesting, but which are not worth the expense, chiefly because the money can be better spent in other ways. It is not easy to determine whether an inquiry comes within one of these three categories or not; and it must be said that, in the present state of scientific and popular criticism in this country, our statisticians receive very little aid in solving the problem. Immersed as they are in practical work, it is greatly to their credit that they keep an open ear for such suggestions as come within their reach, and that they are ready to profit by them.1

In the second place, the vulgar criticism lays no stress on the time and labor and skill necessary for scientifically "working up" the statistical material. It demands quick results. It is impatient of delay. Now, it is of course desirable that the general results of a census, such as the total population and those facts that are of immediate administrative value, should be made public as soon as possible; and that is always done in our census. It has sometimes been said that the tenth census, or at least certain parts of it, were unreasonably delayed, owing to miscalculations as to expense. But the full value of a census is obtained only by careful, elaborate and skilful tabulation and analysis of the raw material. This requires time.

1 The schedules for the eleventh census show very distinctly with what care the officers have studied the experience of the tenth census, of the Massachusetts census of 1885, and of the Interstate Commerce Commission. The superintendent has shown great wisdom in securing the services of such men as Mr. Wm. C. Hunt, Dr. John S. Billings, Mr. Frank R. Williams, Prof. Henry C. Adams, Mr. Geo. K. Holmes and others, who have had previous experience or have made special study of the particular subjects entrusted to them. The need of a permanent census office is most keenly felt when one reflects that it is really only by accident that we enjoy the services of these men a second time.

2 "These delays have, however, been very much exaggerated in popular estimation. The leading reports, those which were specially made up of purely statistical matter, were, with the single exception of that on mortality, published in 1883,-a date which, considering the greater scope of the work, compares favorably with the record of any previous census. The volumes thus published in 1883 were the two volumes of the Compendium, and, of the final quarto reports, the following: vol. i, Statistics of Population; vol. ii, Statistics of Manufactures; vol. iii, Statistics of Agriculture; vol. iv, Statistics of Transportation. The foregoing volumes comprised nearly every