Our Population in 1900

TO forecast for the year 1900 the population of that portion of the earth’s surface now and probably at that date still to be known as the United States of America has been a favorite exercise for our patriotic orators, and even for that austerer race who style themselves “ statisticians.” A few bold spirits have indeed carried these computations unflinchingly out to the middle of the twentieth century, and have gazed full at the intolerable brightness of such figures as 1950 — 497,246,365. There have been Congressmen released from fear, who could contemplate without blinking a population of one hundred and fifty millions on the Atlantic Slope, and two hundred and fifty millions in the Mississippi Valley. But to all fainter souls the close of the century has afforded a natural and easy resting-place in their imaginative flights; and perched on the barrier which divides this much bespattered hundred-years-after-Christ from the next, they have been content with the elevation gained, declining the giddy heights to which so short a continuance, as for twenty or thirty years longer in their ascending course, would conduct them. And therefore it is, and apparently for no other reason, that the popular prophecy of our national growth has stopped at 1900, where, in the gratifying contemplation of a population exceeding that of Great Britain, France, and Germany combined, we have been content to await what the future should bring forth, holding the evil and the good of the century to be sufficient thereunto.

It has not seemed, however, to occur to those of us who have thus indulged in dreams of our national greatness, that if the perfection of the line of population for sixty, eighty, or one hundred years, according to the ratio of past growth, led to a palpable and gross absurdity, suspicion might not unreasonably arise as to the earlier course of that line ; that if causes were certain to operate, at the latest, within the first few years after the beginning of the twentieth century, such causes would probably be felt in some degree, and in an important degree, prior to the close of the nineteenth century ; that consequently if it was impossible that the population should rise by a steady course to be five hundred millions in 1950, it might not be as much as one hundred millions in 1900 ; but, on tlie contrary, it was in the highest degree probable that the great change which was to reduce population from its theoretical maximum as five hundred, to a reality of three hundred, two hundred, or perhaps only one hundred and twenty-five millions at the later date, would be found bringing that population sharply down from its projected altitude fifty years earlier.

As has been intimated, the sanguine view of the national future has not been confined to stump-speakers or Members from Buncombe. It has been put forth officially in more than one census of the United States, with great show of authority, and with precision, not only as to the millions, tens of millions, and hundreds of millions who were destined to inhabit tin’s happy land in 1900, but also as to the hundreds, tens, and units of the fortunate population. Not a man, woman, or child was to be lost through any failure of the statisticians to carry their calculations all the way out, even to the first decimal place. If “the rule of three ” showed that there were to be 100,355,801.6 persons within the United States in 1900, the presumption has, both humanely and patriotically, been taken as in favor of the fractional citizen, and the population at that date been set down at 100,355.802.

The best known of all the definite predictions in respect to the future population of the country are those of Elkanah Watson, who, in 1815, forecast the results of the census from 1820 to 1900. Mr. Watson’s estimates are certain always to be treated with a degree of consideration, from the fact that they were made so early in the history of the country that they were verified with exactness for several successive decades, before the great inevitable change set in. The following are his figures from 1820 to 1860, in comparison with the actual results of the census.

1820 1830 1840 1850 1860
Watson 9,625,734 12,833,645 17,116,526 23,185,368 31,753,824
The Census 9,633,822 12,866,020 17,069,453 23,191,876 31,443,321
Watson's Error -8,088 -32,375 +47,073 -6,508 +310,503

Probably no social philosopher, skilled to discern beneath the variously agitated surface the deep, strong current of human affairs, ever obtained one tithe of the popular applause on account of a prediction fulfilled, which Mr. Watson received, as from decade to decade his estimates of the future population of the United States were thus borne out by the census. At each successive verification, newspaper editors laid down their pens to admire, and took them up again to write, Prodigious ! That a man, a mere human being, should be able to predict fifty years in advance the number of inhabitants in a rapidly growing country, within a fraction of one per cent, seemed to those who knew nothing ot statistical methods, but imagined, as ninety-nine out of a hundred persons did, that Mr. Watson obtained his results by direct and immediate intuition, wonderful, almost beyond belief. And yet, if one will be at pains to examine these much admired predictions, he will find that they were founded upon no prophetic conception of the future, no philosophical analysis of existing forms and forces, nor even upon an exhaustive study of the territorial conditions of the present and future population of the country. There was not even mathematical ingenuity displayed in the computations ; no transcendental processes appear to have been employed ; a school-boy’s arithmetic was sufficient to have carried the scheme out to the end of the twenty-ninth century, as perfectly as Mr. Watson carried it to the end of the nineteenth. Mr. Watson himself assumed no mysterious function in the matter. Plis entire introduction to the estimates, so far as it concerns his method, was as follows ;

‘‘In 1810, it (the population) was 7,239,903. The increase from 1790, the first census under the Constitution, has been about one third at each census; admitting it shall continue to increase in the same ratio, the result will be as follows.”

It will be seen that Mr. Watson’s method was simply to assume an uninterrupted growth of population for. ninety years, and thereupon to compound the population ot ISIO at the rates of increase previously maintained. The whole merit or vice of these predictions was, therefore, to be found in the assumption of an uninterrupted growth. Mr. Watson simply bet nine times upon the red. Five times the red won, — a Wonderful run of luck, certainly ; but when we think that had nullification proceeded, as it was more likelv to do than not, to secession in 1832, the estimate for 1840 would not have been realized; that had not the potato crop failed in Ireland in 1846-7, the estimate for 1850 would not have held good ; that but for the acceleration of European and especially of German immigration between 1850 and 1854, due wholly to domestic causes, the results for 1860 would have been more than a million short of the estimate, we cannot but think that Mr. Watson had a narrow escape on the third, fourth, and fifth ventures. At the sixth, the luck changed. Of the predictions for the three remaining decades of the century, the less said the better ; and as the responsibility for the estimated population of the United States, 18701900, is shared by others, among them two professed statisticians, speaking with the advantage of forty or fifty years of added experience, we shall, at this point, drop all exclusive reference to Mr. Watson, merely remarking that in what we have said, we have intended no disparagement to his eminent services in connection with the industrial development of the country, and no disrespect to his memory. DovTbtless, while he w’as interested in observing the fulfilment of his predictions for three successive censuses (he died in 1842), he would have smiled at the value popularly assigned to the estimates for the latter decades of the century, knowing well that the fundamental assumption was beset by so many chances as to render the remoter results exceedingly questionable.

The late Mr. DeBow, Superintendent of the Seventh Census, one of the most meritorious of the earlier generation of American statisticians, after computing the population of the United States successively on eight “distinct and more or less probable assumptions of future increase,” pronounced the opinion that the figures 100,337,408 “more nearly express the truth than any other for 1900.” [See Compendium of the Seventh Census, p. 130, 1.] The assumption by which this particular result for 1900 was reached would require a population in 1950 of 330,846,389, or an excess of the population of China, according to the better estimates. Whether Mr. DeBow doubted the capacity of the American people to adapt themselves to the use of dogs, cats, and mice, as food, upon so short notice, or for some other reason, he refused the leap, and, like Mr. Watson, stopped short with the nineteenth century.

Precisely what were the data taken, and what the principle of connecting them assumed, in thus-forecasting the probable future population of the United States?

In 1854, when Mr. DeBow made the computations referred to, seven censuses had been taken under the Constitution, with the ascertained population following:

Year. Population. Positive Increase. Increase Per cent.
1790 3,929,214
1800 5,294,390 1,365,176 34.74
1810 7,215,858 1,921,468 36.31
1820 9,600,783 2,384,925 33.05
1830 12,820,868 3,220,085 33.54
1840 17,619,641 4,798,773 37.43
1850 23,067,262 5,447,621 30.92

Such, with the addition of the returns of immigration made to the Department of State, appear to have been the data concerning the population of the United States as a whole which Mr. DeBow used in his computations of the probable increase to 1900. It will be observed that the period 1840 - 1850, the last of the decennial periods in contemplation, had shown a marked decline in the rate of national increase, the per cent gain being but 30.92 against 37.43 for the ten years immediately preceding. A change so marked might not unnaturally have indicated to DeBow’s mind a change in the conditions of population within the United States, and have led him to take a diminishing ratio of increase for the future. But the Superintendent of the Seventh Census wduld seem to have had his own reasons for believing that the causes which effected this falling off between 1840 and 1850 had already done their worst, and to have had no hesitation in assuming the ratio for that decade as the most probable rate for the immediate future. The event proved that, so far as the next succeeding decade was concerned, he was right in not anticipating a further decline in the rate of increase. On the contrary, the Eighth Census found a population of 31,443,321, being a positive gain of 8,376,059, a gain per cent of 36.31.

The Eighth Census brings us to another and professedly an original and independent computation of the population of the United States in 1900. Mr. Kennedy, the Superintendent, was, however, in general only an imitator, and not a successful one, of his predecessor’s methods. In this particular case there is reason to allege something even worse than imitation. In the preliminary report of the Eighth Census, bearing date 1862, Mr. Kennedy presents what purports to be a computation of the future population of the country “ based on the well-known and very correct assumption of a mean annual increase of three per cent.” 1 Treating the ascertained population of 1860 according to this rule, however, we find that in not a single instance does the result correspond to Mr. Kennedy’s table ; and on placing the figures side by side with those of Elkanah Watson, for the first time while writing, we discover, much to our astonishment, that they are identical to the last unit for each decimal period until 1900, and at that point differ only by hundreds in a total of a hundred millions. We now set in comparison the estimates of Watson and DeBow for 1S70-1900, placing opposite their estimates for 1870 the figures of the Census. Here then, at the Ninth Census, we meet the first important deflection from the projected course of population. The ascertained aggregate of 1870 falls short of the estimated aggregate by 3,770.061, according to Mr. Watson, and by 4,255.355, according to Mr. DeBow.

Year. Watson. DeBow. The Census.
1870 42,328,232 42,813,726 38,558,371
1880 56,450,241 58,171,009
1890 77,266,989 79,036,950
1900 100,355,985 100,337,408

From the point of view occupied in either of these computations, there are three ways of regarding this failure of the period 1860-70 to realize the gain in population anticipated therein ; and by consequence three methods of treating the estimated population of 1900. The first is to consider the rapid decline noted in the ratio of national increase as significant in respect to the remaining decades of the century, i. e., as due to causes certain or likely to operate in the future, in an equal, or greater, or smaller degree ; and hence not only to accept the actual loss of the one decade already concluded, but to reduce the estimated ratios for the three unexpired decades. The second method likewise treats the actual loss of the period 1860-70 as irretrievable, but considers it as due to exceptional causes, which have not only ceased wholly to operate for the present, but which are exceedingly unlikely to be again experienced within the century ; and in this view discounts the computed population of 1900 by just the loss realized in the single instance. The third method would be to claim for the country a recuperative power, which will enable it to repair the loss sustained, not only maintaining the assumed ratios in the time to come, but by a display of energy not otherwise to be expected, making good the deficiency of the decade 1860-70, anjl bringing the population of the United States up to a round hundred millions at the end of the century.

These three methods may be discussed in an inverse order. The third is easily dismissed, since it would be in the highest degree irrational in the face of a population in 1870 of only 38,500,000 to predict a population of 100,000,000 in 1900.

The second is the method most likely to receive the countenance of those who have been accustomed to indulge without misgiving in anticipations of an uninterrupted national growth. By what amount, then, must we reduce the final result in 1900, to meet the facts of 1870 ?

It lias been shown that the ascertained population at the Ninth Census was short by 4.255.355, according to Mr. DeBow’s scheme, and by 3,770,061, according to Mr. Watson’s projection. For the farther purposes of this discussion, we will take the mean of these two sums, calling the realized loss of the decade four millions. But this is not necessarily or probably the sum by which the population of 1900 is to be reduced to meet the unexpectedly developed loss of the period 1860 - 70 ; for it is evident that the computations of both DeBow and Watson required that the four million persons thus “turning up missing” in 1870 should have been responsible for a portion of the population of 1900. This they are now ascertained to be disenabled to effect, by reason of their own non-existence at the earlier date. What loss, then, at 1.900 is represented by the loss of four millions at 1870?

It is clear that, without reference being had to the longevity or fecundity of individuals, — a thing wholly impossible, especially as the individuals in this case are not to be found, — the answer to the above question must depend on the answer to the prior question, Out of what class, or classes, of persons, in respect to age, was the loss sustained ? The scope of this inquiry will be most fully appreciated if we make successively four characteristic suppositions. Suppose, firstly, the loss to have been distributed proportionally among all the classes of the population in respect to age : the number of persons, on Mr. DeBow’s computation, who, in consequence of the loss of four millions in 1870 will be returned by the United States Marshals in 1900 as non est inventus, is easily ascertained by “the rule of three” to be 9,770431. Mr. Watson, having pitched the intermediate population at 1S70 somewhat lower than Mr. DeBow, his final term is less reduced by the falling off. Taking the mean of the two, we shall still have, in round numbers, ten millions as the loss to the population of 1900 resulting from the loss of four millions at 1870.

But suppose, secondly, that the loss were wholly out of that class of persons who are in the decline of life. In this case, the loss of such a number of persons would not only not reduce the population of the country thirty years later by a greater number than their own, but would clearly reduce the ultimate population by a number much less than their own, that is, less than four millions, inasmuch as on the one hand, comparatively few of these persons could have been expected, in any probable event, to survive at so distant a date, and on the other hand, by the ordinance of nature, persons of this class cannot be expected to increase the population of the country by offspring ; that is to say, the whole loss at 1870 would, under this supposition, have been out of a class the members of which, as a rule, could not be expected either themselves to survive in 1900, or to be represented at that date by descendants born, or born of parents born, after the present time.

Suppose, however, that the entire loss at 1870 had been out of the class under five years of age: the loss thereby caused to the population of 1900 would have been, not only directly from the loss of those who out of this four millions would naturally have survived thirty years later, but, secondly, from the loss of all the descendants who might fairly have been calculated on as representatives at 1900 of these four million children of 1870. These descendants, however, it should be noted, would generally be in the first degree only, that is to say, the class under five at 1870 would have become 30-35 in 1900, quite too young to have had grandchildren born to them.

But suppose, finally, that the loss at iS7o had been wholly out of the class 20-40 years of age; then the direct and contingent losses to the population of 1900 would have been very much increased, inasmuch as not only would the natural survivorship out of these four millions have been defeated, but also the survivorship out of the children who might have been born to them after 1870, and out of the children of such children ; so that three generations at 1900 would be decimated by the causes which cut down the population of 1870.

Now, it is true that no cause, or combination of causes, could importantly affect by reduction any one of these general classes in respect to age, without appreciably affecting the others. All must suffer with every one, but by no means equally. War affects population differently from pestilence ; the influence of immigration or emigration on the distribution of the population by ages is very marked ; while social habits, going to the birth rate, may cause a disturbance far exceeding that produced by any of the agencies mentioned. It is, therefore, of importance in this connection to ascertain whether the causes that have reduced the estimated population of 1870 have affected the distribution of the ascertained population by ages in such a degree as to materially change the expectation of increase between 1870 and 1900.

Reference to the Table of Ages for the living population at the Ninth Census shows that from each 100,000 of the population there were the following number of persons living within each specified period of life, at 1S60 and 1870 respectively :

Period of Life. Census of 1860. Census of 1870.
0-10 28,665 26,789
10-20 22,524 22,892
20-30 18,211 17,696
30-40 12,789 12,651
40-50 8,314 9,125
50-60 5,043 5,821
60-70 2,827 3,277
70-80 1,109 1,349
80 and over 351 387
Unknown 167 13

Grouping these figures into three grand divisions, we have the fact that, in round numbers, there are 1,500 more persons above fifty years of age, and 1,500 fewer below the age of twenty, in each 100,000 of the population in 1870 than in 1860. On the other hand, the class twenty to fifty holds about the same proportion to the aggregate population as at the previous census. Our space will not serve for anything like an adequate discussion of the degree in which this increase of the aged and sterile class of the population, at the expense of the class under twenty years of age, should affect the growth of population in the next thirty years ; we shall content ourselves with simply pointing out the direction of this tendency. It is at least evident that we must discount the estimated population of 1900 by considerably mere than the 10,000,000 which lias been shown would be the loss at that.date proportional to the developed loss of 4,000,000 out of the aggregate of 1870. This would bring the United States, at the close of the century, distinctly below 90,000,000, — say to 89,000,000, — were all other causes to conspire equally as heretofore to the increase of population.

This last proviso brings us at once to another method of treating the failure of the period 1860 - 70 to maintain the rate of growth characterizing the eight preceding decades of the nation’s history, which is, to regard the relative decline of the last decade as due to causes certain or likely to operate irn the future in an equal, or greater, or smaller degree, and to reduce the estimated ratios for the three unexpired decades of the century correspondingly. If the computations of Watson and DeBow accurately projected the line of the national ascent, according to the rates previously maintained, there was a loss of approximately four millions in the ten years under discussion. To what cause or causes was this loss due ? The natural and immediate suggestion is, of course, the War of the Rebellion ; but can we, on a careful analysis of known facts, maintain the position that the proper effects, whether direct or consequential, of that struggle, bloody and protracted as it was, involved a deficiency of four millions in the otherwise population of the country? This is a question most important in the consideration of the national future ; and while it cannot be answered either way with absolute assurance, reason appears for believing that social forces and tendencies, not heretofore felt, or at least not heretofore recognized, in our national life, are beginning to affect powerfully the reproductive capabilities of our people ; and that these forces and tendencies have contributed in a very large degree within the last decade to bring down the ratio of increase in the native population.

The Report of the Superintendent of Census, November 2t, 1871 [pp. xviii, xix, vol. in population], contains a computation of the effects of the Rebellion on the population : first, through the direct losses by wounds or disease, either during service in the army and navy, or within a brief term following discharge ; second, through the retardation of increase in the colored element, due to the privations, exposures, and excesses attendant on emancipation ; third, through the check given to immigration by the existence of war, and the apprehension abroad of results prejudicial to the national welfare. The aggregate effect of these causes is estimated by the Superintendent as a loss to population of 1,765,000.

There remains but one effect to be ascribed to the war in such a sense, that the war ceasing and the political and social order being measurably restored, further and manifestly new or original effects in the same direction should not be anticipated ; and that is the temporary reduction of the birth-rate consequent on the withdrawal of from twelve to fifteen hundred thousand men from domestic life for an average term of from three and a half to four years. “Speaking roughly,” says the Superintendent, “ one half of these were unmarried men, who on account of their military engagements failed to form marriage relations. The other half were married men whose families were rarely increased by birth during the continuance of the war.” Do we find here explanation of all the loss in population during the decade, not accounted for under the three heads previously mentioned ? This question we can best answer by comparing the number of persons thus withdrawn from domestic life with the total number of the class from which they were taken, and comparing the period during which they were thus withdrawn with the entire term of ten years under discussion. The natural militia of the United States, i. e., the males between eighteen and forty-five, numbered in 1S70, 7,570,487. Taking the middle of the war-period, 1863, the number was probably in the neighborhood of 6,600,000. Assuming therefore the largest number (1,500,000) for the average strength of the two armies, and assuming that this body of men were engaged in military service for the solid term of four years (instead of three and a half), we should still have less than one fourth the natural militia of the country withdrawn from domestic life, and that for two-fifths of the decade ; so that, on these extreme suppositions, the number so withdrawn, taking time into account, would stand to the number not so withdrawn as less than one to nine ; while on the supposition of a smaller aggregate number and a shorter average term, we should reach the proportions of one to twelve, or even of one to thirteen. Inasmuch, then, as births aggregating in the ten years not exceeding eleven and a half millions would have maintained the population of the United States at its numbers in 1861, and have increased that population in the ratio in which it did increase from year to year till 1870; and as this aggregate of eleven and a half million births would have been separated at the latter date by not exceeding eight and a half millions of survivors, it is difficult to believe that the otherwise population of 1870 could have been diminished by this cause to the extent of more than three quarters of a million. Adding this latter number to that number previously given as expressing approximately the losses by emancipation, by the check given to immigration, and by wounds and disease among the soldiers of both armies, we have an aggregate loss to population from the effects of the war, both direct and consequential, exceeding two and half millions. If, then, the probable population of 1870 had been properly projected by the early statisticians of the country, there was a loss of something like a million and a half due to causes other than the Rebellion. If we shall be able to show, or, rather, if a simple appeal to the daily observations of our readers shall suffice to convince them, that these causes are likely to continue and even to operate with increasing force in the immediate future, we shall reach almost an assurance that the population of the United States at 1900 is to be brought down from its projected height as 100,000,000, not only below 90, but even to So, 75, or it may be 70,000,000.

And, indeed, the expectation of the larger result never was a reasonable one, nor could the estimates of Watson and DeBow at any time have been justified by a comprehensive survey of the physical and industrial conditions of the country, or by reference to the experience of any race or people known to history. Geometrical progression is rarely attained, and never long maintained, in human affairs. Whenever it is found, the most improbable supposition which could be formed respecting it is that it will continue. Gibbon has shown that the further conquest is carried, the wider and the weightier become the resistance and the hostility which the conquering power is forced to encounter. So it is with national growth whether in wealth or in population. Not only do the limitations of nature become more and more stringent in reducing the rate of increase, but that increase does of itself create moral and social, not to speak of distinctly political, tendencies. which traverse its own course, and, if not strong enough to defeat further growth or accumulation, do at least m ike every successive gain more slow and painful. It was sufficiently hazardous for Mr. Watson, writing after the Third Census, to predict an uninterrupted and unretarded advance for as many as five decades ; but it was far more lmz; r Ions for Mr. DeBow, writing after he Seventh Census, to predict the continuance of the previous ratio of increase for the remaining five decades of the century ; more hazardous, because the long continuance of that ratio was an argument for and not against a change.

The change came ; came later even than it had been reasonable to expect. It began when the people of the United States began to leave agricultural for manufacturing pursuits; to turn from the country to the town ; to live in upand-down houses, and to follow closely the fashion of foreign life. The first effects of it were covered from the common sight by a flood of immigration unprecedented in history. Even its more recent and more extensive effects have been so obscured by the smoke of war, that the public mind still fails to apprehend the full significance of the decline in the rate of the national increase, and vaguely attributes the entire loss of population to the Rebellion. But a close observer must discern causes now working within the nation, which render it little less than absurd longer to apply the former rates of growth to the computation of our population at 1880, 1S90, or 1900, What rate will be substituted therefor, it would be futile to inquire. As the line of agricultural occupation draws closer to the great barren plains ; as the older Western States change more and more to manufactures and to commerce; as the manufacturing and commercial communities of the East become compacted ; as the whole population tends increasingly to fashion and social observance ; as diet, dress, and equipage become more and more artificial ; and as the detestable American vice of “ boarding,” making children truly “encumbrances,” and uprooting the ancient and honored institutions of the family, extends from city to city and from village to village, — it is not to be doubted that we shall note a steady decline in the rate of the national increase from decade to decade. But it would be merely an attempt at imposture to assume that numerical data exist for determining, within eight or ten or twelve millions, the population of the country thirty years from the date of the last census. As long as one simple force was operating expansively upon a homogeneous people, within a territory affording fertile lands beyond the ability of the existing population to occupy, so long it was no miracle to predict with accuracy the results of the census. But in the eddy and swirl of social and industrial currents through which the nation is now passing, it is wholly impossible to estimate the rate of its progress, even though we may feel sure that the good ship will steadily hold her course, and in time round the point which hopes too fond had — on the strength of a fortunate run made upon a smooth sea, with favoring winds and following floods — predicted would be reached by the blessed year 1900. This much, however, may with diffidence be said: that the best of probable good fortune will hardly carry the population of the country beyond seventy-five millions by the close of the century.

Francis A. Walker.

  1. Report, p. viii.