LARGE-SCALE AND SMALL-SCALE FARMING.*

BY T. N. CARVER, Professor of Economics, Harvard University.

It seems to me that this is one of the most interesting and instructive meetings I have ever attended, and I am particularly glad to hear the fact brought out that an increase in the intensive farming may and does go on while farms are growing a little larger and the farming population a little scarcer. That is, the soil is being made to produce more through more thorough cultivation, not through the use of more labor, but by the use of more capital. That seems to be the significant thing that is happening in the state of Iowa and the agricultural regions of the surrounding states.

It seems to me that in view of the facts that have been brought out today we may be justified in redefining largescale, medium-scale, and small-scale farming. It is obvious that acreage alone is not a basis for this classification. You may have large-scale farming on twenty acres, and you may have small-scale farming on two hundred acres. It seems to me that the definition and the classification should be based on the general character of the business unit known as the farm, with the labor supply as the principal ingredient. Large-scale farming would, therefore, be any kind of farming in which the manager did not do the manual work, but confined himself mainly to the work of superintendence. A twenty-acre market garden worked by a gang of laborers under a superintendent would thus be a large-scale farm. And a medium-scale farm might be called a one-family farm, that is, a farm run by the labor of one family, but a farm on which that labor was equipped with the best teams and the best machinery known to agriculture. That seems to be the type of farming which is developing more and more in the prosperous agricultural sections of the West. Then small

*Remarks at special session on Agricultural Statistics, annual meeting of the American Statistical Association, St. Louis, Mo., December 29, 1910.

scale farming would be of a type which might otherwise be termed peasant farming; that is, where the labor is done by the farmer and his family, but where the acreage is too small to permit of good teams or good machinery. This is not the most productive type of farming, though it is the kind which is being advocated by a good many of our long-distance farmers. They are holding up to our admiration the French peasant. But the French peasant who has to make his living by general farming off a thirty-acre farm must of necessity use inferior methods of production. Such a farm will not support a good team of horses, sometimes not even a yoke of oxen. The farmer therefore finds it more economical to use his milch cow for a draft animal and do the rest of the work by hand. That is, he will have more left for himself by this process than he would if he tried to use a team of horses, for the horses would eat up so large a proportion of the total crop as to leave him less for his own family, even though the horses could do more work than the milch cow. This type of smallscale farming also seems to be disappearing, judging by the statistics which have been given us. It seems to me, therefore, that to the discriminating mind the statistics of population in this agricultural area are things to be thankful for rather than to be alarmed over.

THE CHANGE IN THE PROPORTION OF CHILDREN IN THE UNITED STATES AND IN THE BIRTH RATE IN FRANCE DURING THE NINETEENTH CENTURY.

BY WALTER F. WILLCOX.

The object of this paper is to determine approximately the changes in the birth rates of the United States during the nineteenth century and to compare those changes in direction and amount with the changes which went on in France during the same period.

In European countries as a class the steady and rapid decrease of the birth rate is a comparatively recent fact. If one considers the average birth rate for each ten years of the century or of so much of the century as is covered by the records, the maximum rate is found to have been reached in France as early as 1811-20, in Norway, Sweden, Finland, Austria and Prussia in 1821-30, in Belgium in 1831-40, in Denmark in 1851-60, in Scotland and Spain in 1861-70, in England and Wales, Ireland, Hungary, Switzerland, Germany, Bavaria, Saxony and the Netherlands in 1871-80, in Portugal, Italy, Servia and Roumania in 1881-90. In 16 of the 22 countries named the lowest birth rate was for the last decade of the century.*

The usual opinion of American students, I believe, is that the decline of the American birth rate, about which so much is written and so little known, is likewise a comparatively recent fact. In the valuable paper by Dr. J. S. Billingst the statistical argument was confined to the decade between 1880 and 1890, but the more general discussion with which he concluded shows at various points that he conceived the change to have set in thirty to forty years earlier, or about the middle of the nineteenth century. This conviction has furnished an important support to the argument that the decrease of the American birth rate, starting when foreign immigrants first

Statistique Internationale du Mouvement de la Population, Paris, 1907, p. 141. "The Diminishing Birth Rate in the United States," Forum, June, 1893.

became a powerful influence and manifest where and only where those immigrants settled, was caused by immigration and resulted in little more than a replacement of the earlier by the later stocks.

The belief that the decrease of the American birth rate began not earlier than the middle of the nineteenth century finds support in some conclusions from the figures of the Twelfth Census. In the passage referred to below* I showed that the proportion of children under 5 years of age to each 1,000 women between 15 and 49 years of age during the last half of the century, for which alone the figures were to be had, changed as follows:

This evidence clearly supported the prevailing opinion and led me to the belief that the decrease of the American birth rate began at about the time of the Civil War. I now think this inference was unwarranted and that the decrease began as early as 1810.

The preceding table can be carried back for earlier census years by means of estimates. But for that process it seems wiser to assume the years of child bearing as 16 to 44 instead of 15 to 49, because in 1800, 1810, and 1820 the white women 16 to 44 were reported by the census and estimates are needed only for the colored.

Estimated number of women 16 to 44 years of age.

1900, 1890 and 1880 no estimates required.

1870-women between 18 and 44 given in the tables;

• Twelfth Census: Supplementary Analysis and Derivative Tables, p. 409, ff.

women 16 or 17 estimated by assuming (a) that the white women 16 or 17 bore to the reported number 15 to 17 in 1870 the same proportion (68.1 per cent.) that they did in 1880; (b) that the colored women 16 or 17 bore to the reported number 15 to 17 in 1870 the same proportion (65.65 per cent.) that they did in 1880. By this method 90.1 per cent. of the total number of women was obtained by enumeration and any error in the estimate would be reduced to one-tenth in its effect on the total.

The following table shows the estimated number of women 16 to 44 years of age at each census between 1800 and 1900 with the proportion derived from estimate.

It will be noticed that the largest proportion (35.9 per cent.) was determined by estimates in 1830 and that for earlier dates, notably 1800, 1810 and 1820, the margin of probable error is much less. This is important because it is the general trend rather than the figures for any single year that is important.

Estimated number of children under 5 years of age.

The following table shows in like fashion the number of children under 5 years of age enumerated or estimated for each

census.