An Essay on the Principle of Population

One of the most curious and instructive points of view, in which we can consider lists of this kind, appears to me to be the dependence of the marriages on the deaths. It has been justly observed by Montesquieu, that, wherever there is a place for two persons to live comfortably, a marriage will certainly ensue:
*64 but in most of the countries in Europe, in the present state of their population, experience will not allow us to expect any sudden and great increase in the means of supporting a family. The place therefore for the new marriage must, in general, be made by the dissolution of an old one; and we find in consequence, that, except after some great mortality, from whatever cause it may have proceeded, or some sudden change of policy peculiarly favourable to cultivation and trade, the number of annual marriages is regulated principally by the number of annual deaths. They reciprocally influence each other. There are few countries in which the common people have so much foresight, as to defer marriage till they have a fair prospect of being able to support properly all their children. Some of the mortality therefore, in almost every country, is forced by the too great frequency of marriage; and in every country a great mortality, whether arising principally from this cause or occasioned by the number of great towns and manufactories and the natural unhealthiness of the situation, will necessarily produce a great frequency of marriage.

A most striking exemplification of this observation occurs in the case of some villages in Holland. Sussmilch has calculated the mean proportion of annual marriages compared with the number of inhabitants as between 1 in 107 and 1 in 113, in countries which have not been thinned by plagues or wars, or in which there is no sudden increase in the means of subsistence.
*65 And Crome, a later statistical writer, taking a mean between 1 in 92 and 1 in 122, estimates the average proportion of marriages to inhabitants as 1 to 108.
*66 But in the registers of 22 Dutch villages, the accuracy of which, according to Sussmilch, there is no reason to doubt, it appears that out of 64 persons there is 1 annual marriage.
*67 This is a most extraordinary deviation from the mean proportion. When I first saw this number mentioned, not having then adverted to the mortality in these villages, I was much astonished; and very little satisfied with Sussmilch’s attempt to account for it, by talking of the great number of trades, and the various means of getting a livelihood in Holland;
*68 as it is evident that, the country having been long in the same state, there would be no reason to expect any great accession of new trades and new means of subsistence, and the old ones would of course all be full. But the difficulty was in a great measure solved, when it appeared that the mortality was between 1 in 22 and 1 in 23,
*69 instead of being 1 in 36, as is usual when the marriages are in the proportion of 1 to 108. The births and deaths were nearly equal. The extraordinary number of marriages was not caused by the opening of any new sources of subsistence, and therefore produced no increase of population. It was merely occasioned by the rapid dissolution of the old marriages by death, and the consequent vacancy of some employment by which a family could be supported.

It might be a question, in this case, whether the too great frequency of marriage, that is, the pressure of the population too hard against the limits of subsistence, contributed most to produce the mortality; or the mortality, occasioned naturally by the employments of the people and unhealthiness of the country, the frequency of marriage. In the present instance I should, without doubt, incline to the latter supposition; particularly as it seems to be generally agreed, that the common people in Holland before the Revolution were, upon the whole, in a good state. The great mortality probably arose partly from the natural marshiness of the soil and the number of canals, and partly from the very great proportion of the people engaged in sedentary occupations, and the very small number in the healthy employments of agriculture.

A very curious and striking contrast to these Dutch villages, tending to illustrate the present subject, will be recollected in what was said respecting the state of Norway. In Norway the mortality is 1 in 48, and the marriages are 1 in 130. In the Dutch villages the mortality 1 in 23, and the marriages 1 in 64. The difference both in the marriages and deaths is above double. They maintain their relative proportions in a very exact manner, and shew how much the deaths and marriages mutually depend upon each other; and that, except where some sudden start in the agriculture of a country enlarges the means of subsistence, an increase of marriages must be accompanied by an increase of mortality, and
vice versâ.

In Russia this sudden start in agriculture has in a great measure taken place; and consequently, though the mortality is very small, yet the proportion of marriages is not so. But in the progress of the population of Russia, if the proportion of marriages remain the same as at present, the mortality will inevitably increase; or if the mortality remain nearly the same, the proportion of marriages will diminish.

Sussmilch has produced some striking instances of this gradual decrease in the proportional number of marriages, in the progress of a country to a greater degree of cleanliness, healthiness and population, and a more complete occupation of all the means of gaining a livelihood.

In the town of Halle, in the year 1700, the number of annual marriages was to the whole population as 1 to 77. During the course of the 55 following years, this proportion changed gradually, according to Sussmilch’s calculation, to 1 in 167.
*70 This is a most extraordinary difference, and, if the calculation were quite accurate, would prove to what a degree the check to marriage had operated, and how completely it had measured itself to the means of subsistence. As however the number of people is estimated by calculation and not taken from enumerations, this very great difference in the proportions may not be perfectly correct, or may be occasioned in part by other causes.

In the town of Leipsic, in the year 1620, the annual marriages were to the population as 1 to 82; from the year 1741 to 1756 they were as 1 to 120.
*71

In Augsburg, in 1510, the proportion of marriages to the population was 1 to 86; in 1750 as 1 to 123.
*72

In Dantzic, in the year 1705, the proportion was as 1 to 89; in 1745 as 1 to 118.
*73

In the dukedom of Magdeburgh, in 1700, the proportion was as 1 to 87; from 1752 to 1755 as 1 to 125.

In the principality of Halberstadt in 1690, the proportion was as 1 to 88; in 1756 as 1 to 112.

In the dukedom of Cleves, in 1705, the proportion was 1 to 83; in 1755, 1 to 100.

In the Churmark of Brandenburgh, in 1700; the proportion was 1 to 76; in 1755, 1 to 108.
*74

More instances of this kind might be produced; but these are sufficient to shew that in countries, where from a sudden increase in the means of subsistence, arising either from a great previous mortality or from improving cultivation and trade, room has been made for a great proportion of marriages, this proportion will annually decrease as the new employments are filled up, and there is no further room for an increasing population.

But in countries which have long been fully peopled, in which the mortality continues the same, and in which no new sources of subsistence are opening, the marriages being regulated principally by the deaths, will generally bear nearly the same proportion to the whole population at one period as at another. And the same constancy will take place even in countries where there is an annual increase in the means of subsistence, provided this increase be uniform and permanent. Supposing it to be such, as for half a century to allow every year of a fixed proportion of marriages beyond those dissolved by death, the population would then be increasing, and perhaps rapidly; but it is evident, that the proportion of marriages to the whole population might remain the same during the whole period.

This proportion Sussmilch has endeavoured to ascertain in different countries and different situations. In the villages of the Churmark of Brandenburgh, one marriage out of 109 persons takes place annually:
*75 and the general proportion for agricultural villages he thinks may be taken at between 1 in 108 and 1 in 115.
*76 In the small towns of the Churmark, where the mortality is greater, the proportion is 1 to 98;
*77 in the Dutch villages mentioned before, 1 to 64; in Berlin 1 to 110;
*78 in Paris 1 to 137.
*79 According to Crome, in the
unmarrying cities of Paris and Rome the proportion is only 1 to 60.
*80

All general proportions however of every kind should be applied with considerable caution, as it seldom happens that the increase of food and of population is uniform; and when the circumstances of a country are varying, either from this cause or from any change in the habits of the people with respect to prudence and cleanliness, it is evident that a proportion which is true at one period will not be so at another.

Nothing is more difficult than to lay down rules on these subjects that do not admit of exceptions. Generally speaking, it might be taken for granted that an increased facility in the means of gaining a livelihood, either from a great previous mortality or from improving cultivation and trade, would produce a greater proportion of annual marriages; but this effect might not perhaps follow. Supposing the people to have been before in a very depressed state, and much of the mortality to have arisen from the want of foresight which usually accompanies such a state, it is possible that the sudden improvement of their condition might give them more of a decent and proper pride; and the consequence would be, that the proportional number of marriages might remain nearly the same, but they would all rear more of their children, and the additional population that was wanted would be supplied by a diminished mortality, instead of an increased number of births.

In the same manner, if the population of any country had been long stationary, and would not easily admit of an increase, it is possible that a change in the habits of the people, from improved education or any other cause, might diminish the proportional number of marriages; but as fewer children would be lost in infancy from the diseases consequent on poverty, the diminution in the number of marriages would be balanced by the diminished mortality, and the population would be kept up to its proper level by a smaller number of births.

Such changes therefore in the habits of a people should evidently be taken into consideration.

The most general rule that can be laid down on this subject is, perhaps, that any
direct encouragements to marriage must be accompanied by an increased mortality. The natural tendency to marriage is in every country so great, that without any encouragements whatever a proper place for a marriage will always be filled up. Such encouragements therefore must either be perfectly futile, or produce a marriage where there is not a proper place for one; and the consequence must necessarily be increased poverty and mortality. Montesquieu, in his Lettres Persannes, says, that, in the past wars of France, the fear of being enrolled in the militia tempted a great number of young men to marry without the proper means of supporting a family, and the effect was the birth of a crowd of children, “que l’on cherche encore en France, et que la misère, la famine et les maladies en out fait disparoître.”
*81

After so striking an illustration of the necessary effects of direct encouragements to marriage, it is perfectly astonishing that, in his Esprit des Loix he should say that Europe is still in a state to require laws, which favour the propagation of the human species.
*82

Sussmilch adopts the same ideas; and though he contemplates the case of the number of marriages coming necessarily to a stand when the food is not capable of further increase, and examines some countries in which the number of contracted marriages is exactly measured by the number dissolved by death, yet he still thinks that it is one of the principal duties of government to attend to the number of marriages. He cites the examples of Augustus and Trajan, and thinks that a prince or a statesman would really merit the name of father of his people, if, from the proportion of 1 to 120 or 125, he could increase the marriages to the proportion of 1 to 80 or 90:
*83 But as it clearly appears, from the instances which he himself produces, that, in countries which have been long tolerably well peopled, death is the most powerful of all the encouragements to marriage; the prince or statesman, who should succeed in thus greatly increasing the number of marriages, might, perhaps, deserve much more justly the title of destroyer, than father, of his people.

The proportion of yearly births to the whole population must evidently depend principally upon the proportion of the people marrying annually; and therefore, in countries which will not admit of a great increase of population, must, like the marriages, depend principally on the deaths. Where an actual decrease of population is not taking place, the births will always supply the vacancies made by death, and exactly so much more as the increasing resources of the country will admit. In almost every part of Europe, during the intervals of the great plagues, epidemics or destructive wars, with which it is occasionally visited, the births exceed the deaths; but as the mortality varies very much in different countries and situations, the births will be found to vary in the same manner, though from the excess of births above deaths which most countries can admit, not in the same degree.

In 39 villages of Holland, where the deaths are about 1 in 23, the births are also about l in 23.
*84 In 15 villages round Paris, the births bear the same, or even a greater, proportion to the whole population, on account of a still greater mortality; the births are 1 in 22 7/10, and the deaths the same.
*85 In the small towns of Brandenburgh which are in an increasing state, the mortality is 1 in 29, and the births 1 in 24 7/16.
*86 In Sweden, where the mortality is about 1 in 35, the births are 1 in 28.
*87 In 1056 villages of Brandenburgh in which the mortality is about 1 in 39 or 40, the births are about 1 in 30.
*88 In Norway, where the mortality is 1 in 48, the births are 1 in 34.
*89 In all these instances, the births are evidently measured by the deaths, after making a proper allowance for the excess of births which the state of each country will admit.

Statistical writers have endeavoured to obtain a general measure of mortality for all countries taken together; but, if such a measure could be obtained, I do not see what good purpose it could answer. It would be but of little use in ascertaining the population of Europe, or of the world; and it is evident, that in applying it to particular countries or particular places, we might be led into the grossest errors. When the mortality of the human race in different countries and different situations, varies so much as from 1 in 20 to 1 in 60, no general average could be used with safety in a particular case, without such a knowledge of the circumstances of the country, with respect to the number of towns, the habits of the people and the healthiness of the situation, as would probably supersede the necessity of resorting to any general proportion, by the knowledge of the particular proportion suited to the country.

There is one leading circumstance, however, affecting the mortality of countries, which may be considered as very general, and which is, at the same time, completely open to observation. This is the number of towns, and the proportion of town to country inhabitants. The unfavourable effects of close habitations and sedentary employments on the health are universal; and therefore on the number of people living in this manner, compared with the number employed in agriculture, will much depend the general mortality of the state.

Upon this principle it has been calculated that when the proportion of the people in the towns to those in the country is as 1 to 3, then the mortality is about 1 in 36: which rises to 1 in 35, or 1 in 33, when the proportion of townsmen to villagers is 2 to 5, or 3 to 7; and falls below 1 in 36, when this proportion is 2 to 7, or 1 to 4. On these grounds the mortality in Prussia is 1 in 38; in Pomerania, 1 in 37½; in the Neumark 1 in 37; in the Churmark 1 in 35; according to the lists for 1756.
*90

The nearest average measure of mortality for all countries, taking towns and villages together, is, according to Sussmilch, 1 in 36.
*91 But Crome thinks that this measure, though it might possibly have suited the time at which Sussmilch wrote, is not correct at present, when in most of the states of Europe both the number and size of the towns have increased.
*92 He seems to be of opinion indeed, that this mortality was rather below the truth in Sussmilch’s time, and that now 1 in 30 would be found to be nearer the average measure. It is not improbable that Sussmilch’s proportion is too small, as he had a little tendency, with many other statistical writers, to throw out of his calculations epidemic years; but Crome has not advanced proofs sufficient to establish a general measure of mortality, in opposition to that proposed by Sussmilch. He quotes Busching, who states the mortality of the whole Prussian monarchy to be 1 in 30.
*93 But it appears that this inference was drawn from lists for only three years, a period much too short to determine any general average. This proportion, for the Prussian monarchy, is indeed completely contradicted by subsequent observations mentioned by Crome. According to lists for five years, ending in 1784, the mortality was only 1 in 37.
*94 During the same periods, the births were to the deaths as 131 to 100. In Silesia the mortality from 1781 to 1784 was 1 in 30; and the births to deaths as 128 to 100. In Gelderland the mortality from 1776 to 1781 was 1 in 27, and the births 1 in 26. These are the two provinces of the monarchy, in which the mortality is the greatest. In some others it is very small. From 1781 to 1784 the average mortality in Neufchatel and Ballengin was only 1 in 44, and the births 1 in 31. In the principality of Halberstadtz, from 1778 to 1784, the mortality was still less, being only 1 in 45 or 46, and the proportion of births to deaths 137 to 100.
*95

The general conclusion which Crome draws is, that the states of Europe may be divided into three classes, to which a different measure of mortality ought to be applied. In the richest and most populous states, where the inhabitants of the towns are to the inhabitants of the country in so high a proportion as 1 to 3, the mortality may be taken as 1 in 30. In those countries which are in a middle state with regard to population and cultivation, the mortality may be considered as 1 in 32. And in the thinly-peopled northern states, Sussmilch’s proportion of 1 in 36 may be applied.
*96

These proportions seem to make the general mortality too great, even after allowing epidemic years to have their full effect in the calculations. The improved habits of cleanliness, which appear to have prevailed of late years in most of the towns of Europe, have probably, in point of salubrity, more than counterbalanced their increased size.

1825.

In a census which was made in 1817, of the population of Prussia in its present enlarged state, the number of inhabitants was found to be 10,536,571, of which 5,244,308 were males, and 5,320,535 were females. The births were 454,031, the deaths 306,484, and the marriages 112,034. Of the births 53,576, or 1/8.4 were illegitimate. The proportion of males to females born was as 20 to 19. Of the illegitimate children 3 out of every 10 died in the first year after birth; of the legitimate 2 out of 10.
*97

The numbers here stated give a proportion of births to deaths, as 149 to 100; of births to marriages as 4 to 1; of births to the population as 1 to 23.2; of deaths to the population, of males, as 1 to 33; of females, as 1 to 36; of both together, as 1 to 34½; and of marriages to the population as 1 to 94. The proportion of the excess of the births above the deaths to the population is as 1 to 62; an excess which, if continued, would double the population in about 43 years. As it is not however stated how long these proportions have continued, no very certain conclusions can be drawn from them; but there is little, doubt that the population is proceeding with great rapidity.