cessor, however, and in view of the appropriations, the Director did not feel warranted in materially increasing the rates above those paid ten years ago.

By an amendment to the census act, passed March 24, 1910, the population schedules were made to include an inquiry with regard to the mother tongue of persons born abroad and of foreign-born parents of persons born in this country. This amendment was passed so late that the schedules had already been printed, and some difficulty was encountered in providing for the securing of this information. A considerable number of enumerators failed entirely to ask the question, but it was found possible in the Bureau to edit such schedules, in a large proportion of cases, with almost entire accuracy, by determining the mother tongue from the country of birth. While this, of course, is impossible for certain countries where there is a great variety of races, it can be done with substantial accuracy for other countries, such, for example, as Italy, Norway, and England.

By amendment to the census act, passed February 25, 1910, the Director was required to take a special census of irrigation enterprises. The regular schedules of agriculture used by the enumerators called for the quantity of irrigated land and the crops raised thereon, but additional schedules were prepared for the collection of information directly from the operators of irrigation enterprises. The work of collecting information on these schedules is now in progress and will not be completed until about the close of the calendar year.

FIELD WORK OF MANUFACTURES CENSUS.

Except in a limited number of sparsely settled districts where the enumerators were directed to collect statistics of manufactures, mines, and quarries in addition to the statistics of population and agriculture, the census of manufactures, mines, and quarries was collected through a force of special agents, appointed by the Director of the Census, and entirely independent of the supervisors of the census. These special agents were of two grades, the higher grade having supervisory duties or being employed to canvass the large establishments and the central offices of industrial combinations, while the lower grade performed the greater part of the field work. The number of special agents of the first grade was 79, and of special agents of the second grade 1,132. In addition to special agents of the first grade 106 regular employees of the Bureau were sent into the field mainly for supervisory work.

The census of manufactures, mines, and quarries related to the calendar year 1909. The work of collecting the statistics in the field was begun about February 1 and was nearly completed by the end

of the fiscal year, and since that time has been entirely completed. The returns covered about 345,000 establishments, and the cost of the field work amounted to about $750,000.

The only innovation in the method of collecting the statistics of manufactures for 1910 as compared with 1905 consisted in mailing schedules to the manufacturers shortly in advance of the arrival of the special agent, with the request that they be filled out before his arrival. As was expected, comparatively few manufacturers actually filled out the schedule in advance, but the agents uniformly reported that the schedules served a very useful purpose as an introduction and in familiarizing the manufacturers with the scope and significance of the inquiries.

The schedules of manufactures for the present census were considerably simplified as compared with those previously used. This was especially true of the supplemental schedules used for obtaining detailed information concerning individual enterprises. This simplification of the schedules tended materially to increase the degree of accuracy secured.

By an amendment to the census act, passed February 25, 1910, the Director was required to secure an enumeration of the animals slaughtered for food and the number of hides taken from them. This involved a census of all small butchering establishments, which under the general definition of manufacturing establishments would not otherwise have been included. The work was done by the special agents of manufactures in most areas and by the enumerators elsewhere.

GENERAL ORGANIZATION OF OFFICE FORCE.

The office work in connection with the decennial census has required a very large addition to the force of the Bureau. At the beginning of the fiscal year 1910 the Bureau had on its rolls, as regular employees in Washington, about 650 persons. By June 30, 1910, the force had increased to about 3,000, and by October 31 to about 3,650, consisting in round numbers of 25 administrative officers, 90 special agents, 610 clerks on the permanent roll, 2,750 clerks regularly appointed on the temporary roll through civil-service channels, 115 emergency appointments on the temporary roll, and 65 employees in the machine shop. The latter were, in part, appointed without civilservice examination on account of the necessity of very prompt selections.

The greater number of clerks on the temporary roll were selected in accordance with the general plan provided in the Thirteenth Census act, namely, as the result of a competitive civil-service examination, appointments being apportioned among the States and the selections being made in all cases from the top of the register.

By authority of the provisions of the census act, however, a limited number of persons were employed on emergency appointments to meet the emergency in the rush work of punching population cards. These emergency appointments consisted partly of persons who had had previous census experience and partly of persons selected from the eligible register without regard to apportionment, because of their immediate availability. It is expected to drop this emergency force as soon as the punching work is completed in December.

OFFICE WORK OF THE POPULATION CENSUS.

The office work of the population census consists, in brief, of the following steps: (1) A count of the population direct from the schedules for the purpose of determining the pay of the enumerators and-after subsequent careful examination of the schedules to determine their accuracy-for the purpose of announcing the population of the various localities and States, and ultimately the United States as a whole; (2) such editing of the schedules as is necessary to prepare them for the punching clerks, particularly with reference to mother tongue and occupation; (3) punching of a card for each individual making up the population, showing all the facts appearing on the schedule concerning him, this being done by means of punching machines; (4) the comparing of the punched cards with the original schedules so far as is found necessary; (5) the verification of the cards by means of electrical machines which automatically reject cards in which any of the required holes have not been punched or in which the holes are inconsistent with each other, and the correction of such rejected cards; (6) the sorting of the cards by means of electrical sorting machines into main groups, as determined, for example, by sex, color, or native or foreign birth, several different sortings being required at the different stages of the work; (7) tabulation of the facts with regard to the characteristics of the population from the cards by means of electrical tabulating machines, it being necessary to run the cards through the machines several times in order to take off all of the facts; (8) assembling and publication of the results of the tabulation.

The 300 new electric keyboard punching machines referred to in my previous report, together with the 100 "semiautomatic" electric tabulating machines, were all received from the manufacturers prior to the end of the fiscal year, and most of the punching machines were in operation by that date. Considerable difficulty was experienced at first with the punching machines on account of minor defects in design or in materials or workmanship. Such defects were, perhaps, to be anticipated in a machine based on entirely new principles. The defects have at the present time, however, been

largely remedied by the machine shop of the Bureau, and the punching machines are now working very satisfactorily.

The average number of cards punched per day by each operator on the electric machines is about 1,800 at the present time, as compared with an average of 1,150 punched on the machines of the pantagraph type used in 1900. Because of the difficulties first encountered with the electric machines, about 140 of the old machines are also being used, for day work only. The electric machines are used in two shifts.

Operators punching cards are paid at the rate of 20 cents per 100 for those using the automatic machines and 30 cents per 100 for those using the hand machines. This system of payment is working with entire satisfaction.

The use of the semiautomatic tabulating machines was begun only to a small extent toward the close of the fiscal year, but a large proportion of them are in operation now and are proving most satisfactory. The operators are able, on the average, to tabulate about 16,000 cards per day, as compared with about 7,500 on machines used at the census of 1900. Operators are paid at the rate of 20 cents per 1,000 cards.

Since the close of the fiscal year the Bureau has completed a new automatic tabulating machine of remarkable efficiency. The cards are fed in by machinery instead of by hand, with a consequent increase of about eightfold in speed as compared with the semiautomatic machines. Several other machines of this type are now being constructed and it is expected that a large part of the tabulation will be done by them, with a very great saving in cost.

The sorting machines which are used by the Bureau were bought in 1900 from the Tabulating Machine Company, but have been widened to accommodate the wider card employed, and have been improved in certain other details, thereby greatly increasing their speed. The Tabulating Machine Company brought suit against the Director of the Census, claiming that the alteration of these machines was practically equivalent to the construction of new machines in infringement of the patents of that company. On trial, however, the preliminary restraining order was dissolved and the motion for a preliminary injunction denied. The motion for a permanent injunction has not yet been tried, but the Director of the Census is confident that there has been no infringement of patents.

Great care has been given by the Bureau of the Census to a consideration of methods of presenting the results of the population census, and it is believed that they will be rendered much more valuable to the public than hitherto. An important innovation will be the presentation of the statistics for each county, city, or State in a place by itself. In the volumes of the census of 1900 the facts in

regard to any particular locality were scattered in a large number of different places through several volumes. In addition to this presentation by locality there will be a presentation according to subjects in which the comparative statistics for all cities of 25,000 inhabitants and over and for States will be published. In other words, there will be a duplication of the principal figures in such a way as to facilitate the convenience, on the one hand, of the persons interested in a particular county, city, or State, and, on the other hand, of those interested in any particular subject.

Another innovation will be the presentation of percentages showing the relationship of the figures in practically all cases where such percentages are of any significance. This will greatly facilitate the interpretation and comparison of the statistics.

It is also proposed at the present census to work out from the returns on the schedules statistics with regard to the fecundity of the population, as indicated by the number of children born, and the number of children living, for women of different classes, in comparison with their age and the duration of marriage. A separate set of cards, in addition to those for the population returns proper, is required for this tabulation. A considerable amount of work on this subject was undertaken at the census of 1900, and a large number of cards were prepared, but they were never tabulated and no results were published. I desire to recommend to Congress that the Director of the Census be authorized to tabulate the more important information on this subject for the 1900 census, as well as that for 1910. In due time an estimate of the necessary expenditure will be laid before Congress. This subject is one of profound importance, and the census schedules furnish data by which conclusions of the utmost value can be readily drawn. A plan has been devised by which the expense of punching cards and tabulating the results on this subject for the census of 1910 will be much less than would have been necessary to complete the work on the lines begun in 1900.

OFFICE WORK OF THE AGRICULTURAL CENSUS.

The present census of agriculture is being tabulated by radically different methods from those used in 1900. At that time a system of punched cards and electric tabulating machines was employed. This system, however, was found very expensive and is quite unnecessary in view of the fact that, unlike the population returns, the returns of agriculture need to be presented under only a few broad classifications. In order to secure the arrangement of the facts according to these classifications at the present census, the schedules themselves are sorted, first, according to the color of the farmer; second, according to his tenure, and, third, according to the size of the farm. The data from the schedules for each of these subclasses