The following table gives the diverse results of the several authorities with such detail as the sources admit:

Results of Cuban census of 1817 according to various authorities.

It will be noticed that the maximum and minimum figures for the total population differ by 118,000, or 21 per cent.

Some of the discrepancies in the table may be reconciled. The difference in the number of slaves reported in the first four columns depends upon whether the 25,976 negro slaves said to have been imported during 1817 are or are not included. The 239,830 whites in column 4, when increased by the 19,430 troops on the island and the 32,641 transients already mentioned, give the 291,901 whites of Poinsett. The 154,057 free colored of Poinsett is a clear error, as shown in detail by Humboldt. But why the Habana report of 1882 printed by Behm should give 4,000 more free colored, 14,000 more slaves, and 14,000 fewer whites than the general current of preceding authority can not be explained. On the whole, the figures given by Humboldt are entitled to most respect and to provisional acceptance.

Census of 1819.-The figures given by Behm and Wagner agree with those of Delitsch for 1817, except that the number of free colored is less and the number of slaves greater by 17,058, a change which brings the free colored down to the round number of 97,000.

Census of 1825.-The figures for this census are clearly the roughest estimates. For example, white males, 175,000; white females, 150,000; free colored, 100,000. No confidence can be placed in such results, as the chances against their being the outcome of a count are indefinitely great.

Census of 1827.-Pezuela, in speaking of this census, calls it "the most complete and important that had been published up to that time. It was not, like the previous censuses, a simple summary of population. It was the first description of the public wealth of Cuba, classified both by products and by departments and jurisdictions. It was accurate and satisfactory, free from numerical errors and misleading statements. It showed that in spite of the enormous importations of Africans the whites outnumbered the slaves by more than 24,000. Hence the catastrophe which had overtaken Santo Domingo, where in 1790 there were 15 slaves for every free person, seemed not to threaten Cuba" (see p. 731). The secondary authorities examined and giving the figures of this census, viz, Moreau de Jonnes, Thrasher, the translator of Humboldt, Delitsch, and Behm and Wagner, are in remarkable agreement. While the praise of Pezuela is clearly exaggerated and his word in such a matter is entitled to little weight, yet in default of further evidence the results of the census of 1827 must be accepted as the best possible approximation to the truth. The difficulty in getting access to these early Cuban returns is illustrated by the statement in the Cuban census of 1841 that the writer could not find the original material for the census of 1827, but only the summary of results printed in the Cuadro Estadistico of that year.

Census of 1830.-As none of the more careful authorities, such as Delitsch, Thrasher, Pezuela, and D'Harponville, mention this enumeration, its authority must be rejected.

Census of 1841.-This was an official census for which the original source has been available.

Census of 1846.-The credibility of this count, which reveals an apparent falling off of over 100,000 in the five years 1841 to 1846, depends mainly upon the acceptance or rejection of that of 1841. The internal history of the island affords little ground for believing that the population in these five years actually decreased by more than 10 per cent, a decrease, moreover, confined entirely to the colored population and almost entirely to the slave population, which apparently dwindled by one-fourth in five years. Pezuela, who seems disposed to champion all Cuban figures with more zeal than knowlege, explains the decrease by "the prolonged droughts and disastrous hurricanes which afflicted the country during the five years and destroyed the greater part of the crops." (See p. 732). Delitsch, on the other hand, explains the difference as due to the omission of many slaves in 1846. He declares that there was no actual decrease in the number of slaves in Cuba between 1841 and 1846, and Thrasher disputes Pezuela's explanation, saying that "no great cause, as epidemic or emigration on a large scale, existed during these five years to explain the marked decrease of the slave population; that the material prosperity of the country experienced no decrease except the loss of part of one crop, consequent upon the hurricane of 1845; that church returns of christenings and interments point toward an increase of over 50,000 between 1842 and 1846; and that a capitation tax upon house servants was levied in 1844 and it was generally feared that this mode of taxation would be extended." Hence a powerful motive was at work leading to an understatement of the slave population. Pezuela's explanation is intrinsically improbable because a general cause like drought, hurricane, or epidemic seldom if ever affects a race, much less a part of a race, without also affecting other races subject to similar conditions. On a whole, therefore, the census of 1846, conflicting as it does with that of 1841, can not be accepted as accurate.

Censuses of 1849, 1850, and 1852.—That the figures printed by Behm and Wagner for these years do not rest upon independent counts seems clearly established by the following passage (see p. 732) from the article on population in Pezuela's four-volume Dictionary of Cuba: "At the beginning of 1854, when the task of compiling this dictionary was undertaken, we were seriously embarrassed by the fact that no census or general summary of population had been published since 1846. By the aid of the head of the commission of statistics we were furnished with a number of documents relative to the movement of population-i. e., births, deaths, and immigration-for some years after the last census. These documents enable us to estimate the general population of the island toward the end of 1849."

Census of 1855.—The origin of this return is thus explained by Pezuela (p. 734) in the article already quoted: "From registry rolls and other data found in the statistical office at the end of 1855, Arboleya in his last Manual of Cuba made an estimate of the population." The figures given in Pezuela's table, which follows the quotation just made, agree with those given in the report made to Behm and Wagner in 1882, except that 6,432 persons reported as slaves by Pezuela are transferred by Behm and Wagner to the class of free colored.

Census of 1857.-There is no mention of these figures in Delitsch and no evidence appears that the population of Cuba was actually counted in that year.

Censuses of 1859 and 1860.-These censuses are not mentioned by Delitsch and no reason appears for believing that the figures are the result of independent enumerations.

Census of 1861.-This, like the census of 1841, was a census the official returns of which have been available.

Census of 1862.-These figures are obviously a slightly different rendering of the census figures of 1861.

Census of 1867.-In discussing the figures of this census the scholarly editor of Die Bevölkerung der Erde concludes that they were obtained by combining the returns of 1861 and 1862, and that the returns of 1862 were obtained from those of 1861 by

introducing the births and deaths for the year. At the time that this note was written, in 1874, he concluded that "if we are to hold to anything definite in Cuba we must still go back to the census of 1861," and this conclusion is amply supported by the evidence offered. Delitsch agrees with this rejection of the census of 1867, for, though he includes it in his list, he adds in a note that it rests apparently on an estimate rather than on a new count. For the period after 1861, also, when the most thorough and complete Spanish census of Cuba was taken, it seems best to discard all returns for which the primary sources are inaccessible and which do not appear to have been based upon an actual count of the entire population. Hence the alleged censuses of 1869, 1874, and 1879 may also be rejected as probably mere estimates. The foregoing information regarding the censuses of Cuba gleaned from secondary sources warrants the following conclusions:

1. The only alleged censuses of Cuba taken prior to 1841 which seem worthy of credit are those of 1774 or 1775, 1792, 1817, and 1827.

2. The general consensus of expert opinion regarding these is that they were vitiated by serious errors.

3. These errors, assuming them to have occurred, lay on the side of omissions. 4. The slaves and free colored were probably enumerated with less accuracy than the white population and omissions among them were most common.

5. There is some evidence that the returns for 1792 and 1817 have been changed within the last few decades by a reduction of the reported number of white population and an increase in the number of colored.

6. For 1841 and the years following attention may be confined to those censuses for which official data are available, since none of the others seem to have resulted from actual counts of the population.

7. The censuses of Cuba, therefore, worthy of further attention are those of 1774 or 1775, 1792, 1817, 1827, 1841, 1861, 1877, 1887, and 1899.

As Spanish statistical work was reorganized by a decree of September 3, 1856, some light will be thrown upon the census work in Cuba subsequent to that date by the following extracts from a report upon the organization and progress of statistics in Spain, made to the International Statistical Institute in 1872 by Augustin Pascual, the official representative of the Spanish Government:

"The Spanish bureau of statistics is a division of the ministry of agriculture, industry and commerce, education and public works (fomento). It includes the subjects of the census, statistics, geography, and metrology. Its work is in the hands of an autonomous office, which is organized in two divisions, one for geography, the other for the census and statistics. The latter division collects, coordinates, and publishes statistical returns on all branches of public administration and all subjects which affect the physical, economic, and moral life of the country, or are of scientific interest.

"The general committee of statistics was established in Spain toward the end of the year 1856, and, convinced that its most urgent duty was that of ascertaining the number of inhabitants of Spain, which was at that time unknown, it began its work by a census of population, which was taken May 21, 1857.1 This was a general census of all persons, and taken as of a single day. Every person was entered according to his actual domicile. In order to take the first census of Spain successfully many inquiries were omitted which it might have been well to put, and only those were included which seemed indispensable. Such was the idea at the basis of the instructions dated March 14, 1857. By their terms a personal census was to be taken, the people being classed by place of origin, sex, marital condition, age, and occupation, the other inquiries demanded by modern statistics being left to a subsequent census.

1 1No copy of the Spanish census of 1857 has been found in this country, but we are informed by the secretary of the Royal Statistical Society of London that the work contains no returns for Cuba.

The result was entirely satisfactory to the Government. The number of inhabitants enumerated was 15,464,340. While we do not assume to guarantee the entire accuracy of this number, it was much larger than had been expected. In order to obtain this result all necessary precautions were taken. The clergy, the civil and military authorities, the provincial and municipal corporations, as well as Government agents, vied with each other in performance of their duties. At that time the Government intended to continue similar operations, and when the results of the census were made public by the royal order of September 30, 1858, announcement was made that another census would be taken in 1860 and thereafter every five years. The classification was to be extended as far as possible, indicating both the de facto or present population and the de jure or resident population.

"The royal order of October 30, 1860, decreed a new census, with entry of every name to be taken as of the same date, viz, December 25 and 26, 1860, in all parts of the Kingdom. The population taken was to be only the de facto or present population, because it was believed that the people were not ready to report also their place of residence. The classification of ages was changed. The degree of illiteracy was also to be reported and the classification by occupation included. The inhabitants returned by this census were 15,673,536, an increase of 1.35 per cent in the three years that had elapsed since the census of 1857. This census also gave a classification of the population by occupation, which had not been published in 1857 because of incompleteness in the data. Although the occupation returns in the second census are imperfect, yet they serve as a beginning, the only purpose that the Government had in accepting and publishing them. The census of 1860 was declared official by an order of June 12, 1863, and at the same time a new census was ordered for 1865. It was to include the provinces of America and the Pacific as well as the islands in the Gulf of Guinea. It was also ordered that in the census of 1865 the classification by residence, or the de jure population, should be included. Preparations were begun for the census of 1865, but considering that a census every five years increased the expense and did not produce results much more accurate than those already in hand, since the differences produced in Spain by so short a period were unimportant, and considering further that foreign countries had adopted a ten-year period as the most suitable interval between censuses, the plan for a census in 1865 was abandoned as a result of the royal decree of November 30, 1864, which declared that the next census should be taken in 1870, and thereafter every ten years. By an order of June 7, 1870, a census was decreed for that year, on a date to be fixed later. The census was to be taken name by name and the population to be classified as de facto and de jure. This census was getting under way when the Government, believing that at the close of the last revolution in Spain neither the country nor the administration was in a condition to carry out so great an undertaking satisfactorily, and neither the imperial treasury nor those of the provinces or municipalities could well bear the expense, postponed its execution. It is now (1872) impossible to state when the census will be taken.

"The censuses of Spain taken in 1857 and 1860 were taken upon a day fixed in advance. The population reported was the de facto population. The work was done by means of separate schedules, which were left at the house and later collected by salaried employees. These employees were instructed to answer the questions raised by the head of the family in filling the blanks. The preparation of these family blanks and also of the summaries obtained from them devolved upon municipal committees, and the preparation of blanks for the provinces upon the provincial committees, both of them assisted by salaried employees. The provincial committees made up the résumés or abstracts, and also checked the work of the municipalities and entered their indorsements upon the reports."

The tables in the census of 1860 for all Cuba are reproduced in this volume, pages 710-712. The figures for Cuba bear date March, 1861.

The geographical divisions of Cuba have varied not a little since 1841, and it has been impossible, from the sources and with the time available, to trace the history of these changes in detail. It seems best, therefore, to renounce the effort to present fully the results of the censuses of 1827, 1841, 1861, 1877, and 1887 for any, even the largest, subdivisions of Cuba. In the following tables all results for Cuba as a whole are brought together.

The census of 1841 also gave certain figures for the distribution of population between city and country as follows. The island was divided into 226 towns or urban districts with various names, but grouped together as poblaciones, and 279 rural districts.

The population was then divided as follows: