as thus found is 18.1 years.

That in Cuba is 20.7 years, and in the United States in 1890 it was 21.9 years. The people in Porto Rico, therefore, are more than 2 years younger than those in Cuba and nearly 4 years younger than those in the United States. While the median age in Porto Rico is much below that in Cuba or in the whole. United States, it is higher than that in seven American states in 1890, including most of those with a large negro population. The median age of the several departments of Porto Rico is as follows:

The populations of the 7 provinces differ comparatively little in median age, the range between Guayama and Mayaguez being less than two years, while in Cuba the range between the 6 provinces is over five years, and in the United States the range between the 49 divisions for which the age constitution was reported in 1890 was no less than twelve years. This uniformity in age composition of the several departments of Porto Rico is another statistical witness to the similarity of economic and social conditions throughout the island.

In the preceding table one may notice, however, that the three departments of Bayamon, Ponce, and Mayaguez, containing the only cities of much size in Porto Rico, are the three in which the median age is greatest. It may be that these two facts are connected, that is, that the population of these cities has so high a median age as to raise that for the departments in which they lie. To test the fact the following table has been constructed, showing the median age for these three cities and that for the rest of the department containing the city. The former has been called the urban and the latter the rural population.

From the preceding table it appears that the median age of the rural population in Bayamon, Ponce, and Mayaguez averages below that for the entire island, but that the median age of the three large

cities, Mayaguez, Ponce, and San Juan, is from three to five years higher than that for the rural districts in the same department Hence the hypothesis which the table was designed to test is established. The population of these three cities treated as a unit has a median age of 22.2, while the median age for the rest of Porto Rico is only 17.7 years. The average difference between city and country in Porto Rico is more than double that between the two most widely divergent departments.

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. The table states the proportion that the number of persons in each ten-year period from the begin ning to the end of life made of the total population of all ages.

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

This table shows that the proportion of children under ten in Porto Rico is notably higher than it is in Cuba or the United States. It is also much higher than in most countries of Europe, for the only European country with a higher proportion among the eighteen mentioned in an accessible summary of recent censuses1 is Bulgaria, in which 31.5 per cent of the population was reported by the census of 1888 as under ten. There is no American state in which, by the census of 1890, the proportion of children under ten was as large as it is in Porto Rico. If the population of the United States be taken as the standard, the population of Porto Rico contains many more young children, a few more children in their teens, about the same proportion of persons in the twenties, but a noticeably smaller proportion of persons in each subsequent decennial period until the ages of eighty and over, when the proportions are apparently about the same as in the United States, or even greater. Of persons over sixty there are in Porto Rico only 40 in 1,000, while in the United States there are

1 Allgemeines Statistisches Archiv III, 472 (1894).

62. If we assume that in both countries children under ten and aged persons over seventy are supported by persons between ten and seventy, then in Porto Rico there would be 47 such dependents, young or old, for every 100 persons between ten and seventy, while in the United States there would be only 36.

There are fewer elderly persons over sixty in Porto Rico than in the United States or Cuba, and all three have a lower proportion of aged persons than any of the eighteen countries of Europe. The small proportion in Porto Rico is due in part to the rapid growth of population, but also in large degree to the short average lifetime of the people, resulting from unsanitary conditions, ignorance regarding care for the health, and poverty, all of which are widely prevalent among certain classes.

The analysis may be carried one step farther by finding the proportion of the population belonging to each period of five years between birth and death. The results, in comparison with those for the United States and Cuba, and also with an artificial stationary population from which the errors due to misstatement of age or to uneven growth of population have been excluded, are contained in the following table: Age constitution of the population of Porto Rico compared with that of Cuba and the United States, by five-year age periods.

The great number of young children previously noted is emphasized by the preceding table. In each of the first two age periods there are about 3 more children to 100 of the total population than there are in the United States; that is, in an average 100 Porto Ricans there are between 6 and 7 more children under 10 than there are in an average 100 Americans. Porto Rico has a larger proportion of children under 5 than any American state in 1890, and a larger proportion of children between 5 and 10 than any American state in 1890, except 8490-00

4

South Carolina and Mississippi. This large proportion of young children witnesses clearly to a very high birth rate on the island, and therefore, as the population has not increased with very great rapidity and has not been much influenced by immigration or emigration, it testifies indirectly to a very high death rate.

An examination of the preceding table shows also that during adult life, and especially the later years, the proportions, and therefore the numbers, in Porto Rico belonging to the successive quinquennial groups vary irregularly. For example, the proportion of persons between 50 and 55 is much larger than the proportion of persons between 45 and 50. One would expect the number in each group through middle life to fall below that of the next younger group by a somewhat constant proportion. That it does not is perhaps sufficiently obvious from the table, but the fact is more distinctly brought out in the following derivative table:

Number and per cent by which the reported population at the age group named fell below the number in the preceding age group.

Such irregularity in the decrease with advancing years is counter to all the probabilities in the case. The most simple hypothesis that arises to explain it is errors in the reporting of ages. Where such errors occur they reveal themselves in the large number of persons whose age is reported as a multiple of 5 or especially of 10. Hence, as a result of this tendency, quinquennial groups containing a multiple of 10 are erroneously swollen and the intervening groups correspondingly diminished. An examination of the preceding table will show that this is true of the reported ages in Porto Rico. Further evidence of the irregularity may be found in the following table. The number of persons in each quinquennial group has been compared with half the sum of the numbers in the groups immediately preceding and following. If the curve representing the population by age groups were a straight descending line (for a stationary population and through the