is the science (or, still more strictly, the art) of applying those practical rules, which must be followed, in order to insure a right direction of such attempts. As these rules are daily better understood, it is difficult to say what serieses of cause and effectwhat classes of social phenomena-may not be eventually brought within its range. Nevertheless, it is true that a great distrust is felt by most observers, of the value of those Tabular results in which statistical information is generally conveyed. So far as regards particular tables, this want of confidence in the conclusions drawn from them is generally well founded enough. Very few statistical writers have the patience requisite to examine sufficiently the accuracy of those numerical statements which form the bases of their reasonings. Still fewer have shown the necessary caution and sagacity-we might almost add the necessary modesty-by allowing sufficiently for their own ignorance, and for the great imperfection of the best data which we possess; when they have endeavoured to apply their acquired knowledge to the discovery of principles, The zeal with which they analyze the complicated appearances which society presents, and their eagerness to seize on the first analogies which present themselves on the surface, often remind us of the enthusiasm with which the surgeons of early times, when anatomical science was in its infancy, sought for the principles of thought and life in the dead body which they had scarcely skill enough to dissect. But the distrust to which we have alluded, is a distrust of the method itself, and not merely of the success with which it has hitherto been pursued. In point of fact, habituated as we are to consider the movements and actions of human beings as the result of their own free agency-regarding every individual as a microcosm, a creature of impulses and habits, partly, indeed, determined by circumstances, but still mainly his own, and from whose conduct, under given conditions, it appears almost impossible to conclude with any high degree of moral probability as to the conduct of another-many are apt to regard as a mere chimera, the notion of arriving by numerical calculations, at results sufficiently regular to afford data for reasoning on the conduct of thousands and of millions. And yet the more we examine the subject, the more certainly do we discover that the same rule prevails in moral as in physical phenomena-that we can fix the probability of a particular event, in one case, from investigating a number of similar past events. To take an example familiar to every understanding. If we had no materials for comparing the value of different lives-in other words, if the statistics of life were unknown to us-it would be obviously impossible to form any calculation whatever of the probable length of an individual life. But by examining the particulars of a great many cases, we arrive at conclusions sufficiently accurate to influence our conduct, and are enabled to subject what is roughly called accident, or destiny, to general rules of calculation. The life of one man is liable to a thousand contingencies which mock our powers of divination. Compare a thousand more lives similarly circumstanced, and the influence of contingencies seems to disappear before that of general laws. The case is precisely the same with those effects of which the proximate cause is the free will of man. Nothing at first sight seems more arbitrary or uncertain than the course which any one man will pursue, where circumstances, so far as they are known to us, do not seem to act with any compulsory force on his judgment. Take ten-one hundred-or one thousand men, whose choice is made under similar circumstances; and the greater the number of individuals compared, the more does the slightest pressure of external influence-the mere balance of motives-seem to amount to an irresistible force, effacing all varieties of human choice or caprice. The results of an individual will seem to disappear, it has been well said, before the mean results of innumerable wills: in other words, under the weight of the vast machinery of moral causes; and differences of temper and disposition sink into mere modifications of general laws, subject to calculation equally with those laws themselves. And thus, a first acquaintance with the very striking deductions which are actually obtained by the mere numerical display of facts, relating to the moral condition of society, is apt to change this common distrust of statistical investigation into overconfidence. There is no study in which the eager enquirer is more ready to over-estimate, we will not say the value of the method he pursues, but the actual proficiency which he has made in it. It would seem,' says M. d'Angeville-observing on the extraordinary fact, that not only the number of murders committed, but the proportion between the different means which are employed for the commission of murder, scarcely varies to any important degree, from year to year, in so extensive a country ast France- as if the free will of man existed only in theory, and as if every society contained within its bosom germs of evil ' which must infallibly develope themselves. The harvest of the fruits of the earth is more subject to variations than the har' vest of crime!' It is plain, then, that by statistical calculations we obtain a much greater uniformity of conclusions, in our enquiries respecting the moral condition of a country, than those who have not studied the subject are apt to imagine. To take, for example, the instance of crime already alluded to, as one of the most popular and easy instances of the employment of this mode of enquiry the number of crimes committed in extensive districts, the proportion of criminals of different ages, the proportion of crimes with and without violence, and so forth, are found to vary so little from year to year, still less in periods of years,-allowance made for the movement of population,-as to prove that some very extensive and general causes produce these phenomena. The next and most important question is, what assistance will statistical enquiry afford the political philosopher in the search after these causes ? In order to answer this question, it may be worth while to examine the principles laid down by M. Quételet-the most philosophical of writers on statistics-as to the degrees of difficulty which attend the application of this method to different subjects. We have taken the liberty of somewhat abridging a passage from one of his Works. In recapitulating what has been said as to the possibility of measuring those qualities of men which are appreciable by their effects, it appears to me that numbers may be employed without absurdity in the following 'cases: * 1. When the efforts can be estimated by means of a direct measure which can show their degree of energy; as those produced by the application of strength, swiftness, or activity to employments of the same description;-perhaps, also, the application of the faculty of memory to similar objects. † 2. When the qualities are of such a nature that the effects are always nearly similar, and the degree of the quality only depends on the frequency of the effects: such as, the fecundity of females, drunkenness, &c. If two men, placed in similar circumstances, get regularly drunk, the one twice a-week and the other once, their relative propensity to drunkenness is as two 'to one. 3. Lastly, numbers may still be employed, in cases where the causes are such that both the frequency and the intensity of the effects must be taken into consideration: although the difficulties then become very great, and even insoluble in several cases, from the small amount of data which we possess at present.' For example, in determining the comparative moral condition of two districts or places, it is necessary to possess an account not only of the relative number of legal offences, but of the character of those offences-the comparative efficacy of the administration of justice, since, where law is ill administered, few offences. are prosecuted the comparative strictness of police-the comparative number of second or reiterated offences, since it is necessary to ascertain what proportion of the population is conversant with guilt-and many other circumstances which will suggest themselves to the reader versed in this species of enquiry. When all these are ascertained, the comparative amount of morality (as far as regards the commission of offences against penal laws) is determined; and then, and not until then, the enquirer is prepared with data on which he may proceed farther in searching after the primary causes which may have influenced that amount in different countries,-difference of race, religion, government, education, and so forth. For our own parts, therefore, instead of regarding with distrust, or treating with ridicule, information on such subjects. conveyed in the form of numerical statements, as a fallacious. source of reasoning, we confess that we are better pleased with the aspect of a work full of tables, and perplexed with intricate arithmetic on the moral or physical state of a people, than with the profoundest a priori speculations-provided only that we have tolerably good confidence in the accuracy of the numbers themselves. The more these calculations are multiplied-the more attempts are made to bring minute and what are commonly termed accidental peculiarities of society within the scope of numerical estimates-the more materials are accumulated for the use of social philosophy. Results, seemingly the most insignificant, and collected, perhaps, with no view to immediate use, may prove of service, at some future time, in illustrating principles and in correcting partial deductions drawn from other serieses of facts. Such is the value of the very laborious and unpretending work which is now before us. The author does not profess to investigate (except incidentally) the causes of the various social phenomena which his country presents; but he has endeavoured to estimate the physical and moral circumstances of its population, by a comparison between the state of the several departments, drawn chiefly from official sources. In this synopsis. he has included various particulars which have not usually been made the subjects of statistical comparison. As he has explained at full length the data on which each comparison proceeds, the reader is enabled to judge for himself the value to be attributed to it; and if here and there these data appear to him insufficient, it will thus be his own fault if he builds on them a greater amount of conjecture than they will fairly support. The tables contained in this work, and the maps which accompany it, are constructed on the principle adopted by M. Charles Dupin, of indicating different degrees (as of density of population, crime, pauperism, and so forth) by tints of different depth. A portion of them presents a comparative view of the movement of population in France; for which the details lately published by the Ministry of the Interior afford ample materials. Another portion relates to the physical condition of the people, as evinced by the stature and the constitution of the recruits in each department; a curious subject, for which the means of investigation are afforded by the French law of conscription. According to the present regulation, the minimum height of a soldier is fixed at 1 mètre 56 centimètres-something less than 5 feet 2 inches English; consequently, the proportion of recruits refused by the examiners, by reason of deficient stature, to those admitted -together with a computation of the average stature of the young men drawn in each department for the conscription-afford pretty satisfactory data for estimating the condition of the inhabitants of each district in this important respect. It appears from this investigation, that, short as the stature of the French race generally is, it is most remarkably so in all the western and central departments, and the greater part of the south; in which from 300 to 800 exemptions are annually pronounced for 1000 recruits obtained; in other words, from a sixth to two-fifths of the young men of twenty do not exceed the height of 5 feet 2 inches-a population scarcely to be matched, out of Lapland, for diminutiveness on the other hand, the men of the northern and northeastern departments appear to reach the full average of the European stature. In the next table, that which contains the result of the examination of recruits as to their physical constitution, M. d'Angeville has arrived, to his own satisfaction, at the singular conclusion, that the shortest races are almost invariably the most robust, or, at least, subject to the fewest defects. of constitution. This he grounds on the fact, that the exemptions for defective constitutions are most numerous where those for defective stature are fewest. But we suspect he has fallen into a very obvious mistake. He admits, what, indeed, no one can well doubt, that out of a given number of individuals taken by chance, those above the medium height are also in general more robust than those below it. Now, it is plain enough that the regimental examiners apply first the test which gives least trouble; namely, that of stature ; so that, wherever the greatest number of undersized recruits are rejected, a smaller proportion -- |