males in that city. The foregoing table, however, shows that the proportion of males in the province of Habana, outside the city, was somewhat greater than the proportion in the city itself. It may be worth while, therefore, to see whether other cities had a lower proportion of males than the rural districts. The following table gives an answer to the question. The 14 cities referred to include all those having a population exceeding 10,000, together with the city of Pinar

The excess of males in Cuba is thus seen to hold only in the rural districts. In the cities there were nearly 10,000 more females than males, but in the country about 67,000 more males than females. In an average group of 1,000 city folk there were 20 more females than males, but in an average group of 1,000 country folk there were 62 more males than females.

In the following table the distinction between urban and rural population has been extended to the provinces, and for purposes of simplicity only the columns for males have been retained:

The difference between city and country in all other provinces is several times as great as it is in Habana, and rises to a maximum in Puerto Principe, where in every 100 country residents there are 10 more males than there are in the capital city of that province. It is in Habana province alone that males outnumber females in the cities. Elsewhere they are in a decided minority. This difference may plausibly be connected with the large number of immigrants in the cities of that province, notably in Habana. In every one of the 14 cities separately returned, except Habana and its suburb Regla, the females outnumber the males.

AGE.

(See Table IX.)

CUBA AS A WHOLE.

Probably the best single and simple expression for the age of a great number of people like the inhabitants of Cuba or the United States is what is called the median age; that is, the age such that half the members of the population group under consideration are younger and half are older. To compute it accurately the census tables should present the ages by single years. That information being given, it is easy to ascertain within what single year of life the median age must lie. It is then assumed that within the year of age thus fixed the persons were evenly distributed; in other words, that there were as many persons living in the first tenth of the year or the first month as in each other tenth or month. In this way the median age of the population of the United States in 1890 has been fixed at 21.92 years. The present census of Cuba reports ages not by single years, but, in most cases, only by five-year periods.' Hence to get the median age it has been necessary to distribute the population of Cuba in a single fiveyear period to the several years. For this purpose it has been assumed that the number at each year of age in the five-year group bears the same proportion in Cuba, as in the United States, to the total for the five years. Thus the median age in Cuba has been found to be 20.78 years. That in Porto Rico is 18.18 years. The people of Cuba, therefore, were more than a year younger than those of the United States, but more than two and a half years older than the people of Porto Rico.

The median age is a summary expression of the age constitution and gives only a preliminary idea of the facts. The analysis is carried a step farther by the following table, in which the three population groups are compared in more detail. Here and elsewhere no comparison is made with Spain because of the meager statistical information about that country. The table states the proportion that the number of persons in each ten-year period from the beginning to the end of life made of the total population of all ages.

'The division of the group 15-19 at 17 enables one to know the population of school age, 5-17, and that of age to serve in the militia, 18-44. The division of the group 20-24 at 20 enables one to know the potential voters. The division of the group 0-4 into single years allows a study of the balance between birth rate and death rate during the early years before it is seriously affected by migration.

Age constitution of the population of Cuba compared with that of Porto Rico and the United States, by ten-year age periods.

This table shows that the proportion of children under 10 in Cuba was less than in the United States and much less than in Porto Rico. On the other hand, the proportion of persons between 10 and 20 was much higher and that of persons between 20 and 40 somewhat higher than in either of the other countries. The proportion of persons in Cuba between 40 and 90 was somewhat less than in the United States, but, with a slight exception for the last ten years, greater than in Porto Rico. The proportion beyond 90, which was larger than in Porto Rico or the United States, points not to a greater proportional number of very aged persons in Cuba, but to greater errors in the returns, whereby the true age has been exaggerated. If the age composition of the population in the United States be taken as a standard, there were in Cuba few children, many youth, an average number of young adults, and a small number of persons who had passed the meridian of 40. An accessible summary giving the proportion of children under 10 and of adults over 60 in 18 European countries at the last censuses shows that Cuba had proportionally fewer children under 10 than 14 of these countries, but a larger per cent than Belgium (22.4), Switzerland (21.7), Ireland (20.8), or France (17.5). The per cent of persons over 60 (4.6) was lower than in the United States (6.2), and that was lower than in any of the 18 countries of Europe. The small proportion of aged persons in the United States may be explained by the rapid growth of its population; but in Cuba, where the population has increased only 4 per cent in twenty-two years, the cause must be sought rather in unsanitary conditions, ignorance regarding care of the health, and poverty, all of which are widely prevalent among certain classes on the island and result in a short life.

The analysis may be carried one step farther by finding the propor

1 Allgemeines Statistisches Archiv III, 472 (1894).