The Brussels classification, embracing in all only 186 items, is necessarily general in character. While fairly adequate in regard to raw materials, important details are lacking in the case of manufactured goods. For the various textile manufactures only yarn and fabrics are distinguished. Machinery has only eight subdivisions. Other manufactures of iron and steel are combined in a single item. Manufactures of india rubber are all united. Without doubt the need for an elaboration of the schedule will appear after it has been put in force.

Viewed, however, merely as a first step toward uniformity, the adoption of the Brussels classification offers distinct advantages both as a means of improving commercial statistics and as tending to bring about closer conformity in customs matters. Each of these points calls for separate treatment.

That the statistics of foreign trade of different countries are far from being comparable is generally recognized, but few realize the extent to which such incomparability exists. Of the various causes responsible for this condition, differences in classification account for the most serious apparent discrepancies.

Differences in classification. The primary basis of classification is sometimes found in the origin of the product or its degree of manufacture, as in France; sometimes, as in most South American countries, the raw materials and their various manufactured forms are grouped together in great classes, such as "textiles"; sometimes, as in Germany, both these divisions are made; sometimes the articles are arranged in alphabetical order. The component material determines the classification of articles in some tariffs, while in others regard is had to the use of the article.1 Commercial statistics follow the tariff in practically all countires and follow it slavishly where commercial interests are not sufficiently organized to demand additional details. Statistical classification is so intricate and so varied that international comparisons are often impossible and always liable to error.

A most striking illustration of apparent discrepancy in statistics due to differences in classification is found in the shipments of pumps from the United States to Cuba. In the fiscal year ending June 30, 1914, the American returns recorded a value of $291,350 for pumps and pumping machinery" destined to Cuba; while the Cuban returns show imports from the United States of pumps worth only $19,631. A careful examination of the Cuban tariff and customs decisions brings out clearly the reason for the apparent discrepancy between the two sets of statistics. The item described as "pumps" in the Cuban statistics doubtless refers to the phrase "steam pumps" in tariff No. 218. But even these, if intended for use on farms or in sugar factories, are admitted at greatly reduced rates of duty under the head of "Agricultural machinery and implements" (No. 216) or "Machinery and implements for sugar plantations" (No. 215). The vast majority of the pumps imported are undoubtedly declared for customs purposes under one of these two classes and appear under the corresponding item in the statistics. Pumps other than steam pumps not intended for use on farms or

1 Some variations in the classification of typewriter ribbons are pointed out in Foreign Import Duties on Office Appliances (Bureau of Foreign and Domestic Commerce, Tariff Series No. 29), p. 10.

plantations are classified for duty purposes and undoubtedly for statistical purposes under tariff No. 226, "Machinery and apparatus not specified.'

This instance should not be construed as reflecting in any manner on the accuracy of the Cuban returns. Without doubt numerous cases could be found in the statistics of the United States where, owing to differences in classification, some articles would appear to be understated in comparison with the returns of other American countries. The illustration from the Cuban statistics has been cited merely to emphasize the need of greater uniformity in classification, if general readers are not to be misled.

The adoption of a uniform classification would do much to make the statistics of different countries more comparable. Still other means to this end would be found in international agreement upon the definition of "imports" and "exports," upon methods of determining weight and value, and upon the definition of "country of origin" and "country of destination." These points can well be left to the future, after united action in respect to classification has proved its success.

Varying definition of "imports" and "exports."-The American statistical system is based on the British system. Exports are divided into two classes, according as they represent domestic products or products of foreign countries. Imports represent all goods landed, whether for immediate consumption or for warehouse; only goods for immediate transshipment or for direct transit through the country are excluded. In most continental countries, on the other hand, both imports and exports are divided into "special commerce" and general commerce.' General commerce includes all goods landed and all goods shipped, usually with the exception of the direct transit and transshipment trade, although in some countries these likewise are included. Special commerce represents goods imported for consumption within the country itself and exports produced within the country. Imports of raw materials for reexportation in manufactured form are in some countries included and in some countries excluded, and similarly the articles so manufactured and reexported are included by some countries and excluded by others.

These differences, of definition account for some of the apparent inconsistencies in international statistics. As a rule, however, the discrepancy is not nearly so great as results from some of the other causes cited.

Different method of determining weight and value. While in general the weight shown statistically is intended to represent net weight, there are numerous exceptions to this rule. In the case of value, the resultant discrepancies in commercial statistics are much more apparent. In the majority of countries the effort is made to show the value of exports when they leave the country and the value of imports when they reach the customs frontier, inclusive of freight and insurance, but exclusive of import duty. The United States is an exception since, in the case of imports, the foreign value (f. o. b. port of lading) instead of the value on arrival (c. i. f.) is prescribed by law. Freight charges are not, as a rule, sufficient to cause a marked discrepancy, but the exclusion of certain nondutiable charges, such as export duties in the country of origin, make our statistics of,

say, imports of Brazilian rubber, incomparable in point of value with the British statistics of imports of the same commodity.

The method of determining the value in the United States, the United Kingdom, and British colonies is by the declaration of the exporter or importer supplemented in the case of imports by the invoice. In most of the continental countries, on the contrary, the importer or exporter is required to declare only the quantity of his shipment, while the value is calculated according to a list of average prices fixed annually by a board of experts. This latter system is a gradual development from the old system of official values, when ad valorem duties were imposed according to the legal value, which remained unchanged for a long period. That method of assessing duties is still in force in some of the South American countries where, as a consequence, the unit value shown in the import statistics is constant as long as the valuation tariff remains in force.

Inaccurate declaration of country. The exports of raw cotton from the United States to Germany in the calendar year 1913, according to our own returns, amounted to 2,658,640 bales of 500 pounds each, while the German returns for the same year showed an importation from the United States of only 1,628,828 bales. On the other hand, according to our returns, 117,662 bales of 500 pounds each were shipped to Russia, while the Russian returns show 342,326 bales imported from the United States. In the case of Germany we exported a million bales more than the German returns record, while Russia credits this country with more than twice the amount that our statistics show. The large transshipment trade through the free port of Hamburg and other German ports explains the first of these discrepancies. Shipments declared for Hamburg, without indication of any further destination, are of necessity recorded in our returns as exported to Germany, while in the German returns only the cotton entered for consumption is recorded. In the case of Russia, the reason for the discrepancy is much the same. Our commercial relations with that country are largely indirect, and as a result large quantities of American cotton for Russian consumption are shipped through other European ports, largely for account of firms located in such ports. Such shipments have to be recorded in our returns as exported to the countries in which the transshipment ports are located. If in the Russian returns all the cotton imported could be traced back to its country of origin, the discrepancy would undoubtedly be much larger.

As a general rule, import statistics are more trustworthy than export statistics. In the case of exports, the ultimate destination. of the goods is frequently unknown; while in the case of imports the country of origin, although sometimes difficult to trace, can usually be ascertained if a sufficient effort is made.

Without question uniform customs practice and uniform tariff classification are of far greater practical importance than any improvement in statistical methods. In 1911, before the Pan American Commercial Conference, I advocated a simplification of tariff systems and a uniform customs classification for all American countries (proceedings, p. 208). It is unlikely, however, that any such plan can at once be introduced. Every nation is intensely jealous of its tariff independence. The ideal of a uniform customs classification can most surely be realized, I believe, indirectly, through a standardization of the statistical schedule.

Commercial statistics and the tariff are vitally related. An instance has been cited where their relation caused a serious statistical discrepancy. A marked improvement in statistical classification should, on the other hand, react favorably on the tariff classification. Statistical returns are collected in the regular course of customs procedure by customs officers who must be familiar with the statistical schedule. When the returns are published so as to facilitate international comparisons, the customs practice of different countries will be more easily learned. The tendency toward uniformity in tariff classification will be promoted if the international returns show the duties collected on each article as well as the values and quantities of imports and exports. Excessive duties on any article, compared with those levied by neighboring countries, will stand out, and strong arguments for their reduction may be drawn from such comparisons.

It is coming to be generally admitted that the tariff question has two aspects, one of policy and one of technique, and that the best results can be attained when the legislature confines its action to the fixing of rates, while leaving to administrative officials all matters of technical detail. Commercial statistics, when presented according to a uniform classification, and the advice of experts in the customs service can scarcely fail to leave their impress, in the way of closer international uniformity, on the wording of any new tariff law adopted by one of the contracting States.

The consent of the various Governments to any projected reform is not enough. To be successfully and continuously applied the measure must be actually supported by the appropriate officials. LatinAmerican statisticians are generally progressive men, with a broad international point of view. They can be depended upon to carry out with intelligence and energy the uniform classification proposed and to work for its gradual extension, with the necessary subdivisions, to customs practice.

(1) Report of Surgeon General Rupert Blue on the Standardization of Sanitary Regulations on Imports.

SIR: In compliance with your letter of July 21, 1915, I have the honor to submit herein a report containing a tentative plan for the certification and treatment of freight from or via South American ports destined to United States ports.

While your letter does not indicate that any sanitary regulations, save those concerning imports, are required, I feel that it would be impracticable to suggest a set of regulations which would apply effectively to imports and be economical of administration which did not include certain necessary measures regarding the other phases of maritime quarantine. For example, any kind of cargo, even foodstuffs, coming from a port not infected with plague would be safe; but a good many people believe that because there are no human cases of plague in a port that said port is free of plague, notwithstanding the fact that the rats may have the disease and may enter a vessel either independently or along with the freight. Therefore any set of regulations governing freight inspection and certification, so far as it applies to plague, must, to be effective and economical, deal with the human element in the situation.

Freight from a cholera-infected port might be absolutely safe so far as transmitting the infection of cholera is concerned, but if the water delivered to the vessel was not pure, and if there was not a guaranty that none of the personnel of the vessel were cholera carriers the freight regulations would really be of no consequence. That is to say, the certification of freight must include the consideration of the rat and rat flea in plague; the body louse in the case of typhus fever; the mosquito in the case of yellow fever; and the cholera carrier (man) in the case of cholera.

Therefore, in preparing this report to you I propose to make it in a condensed form, paragraphed, and to touch not only on those factors which at first may appear to be outside the scope of freight inspection, but I shall also make suggestions as to the easiest way in which the most favorable conditions in the Pan American ports can be obtained in the manner most acceptable to everyone concerned. These suggestions are as follows:

TENTATIVE PLAN TO FACILITATE THE IMPORTATION AND EXPORTATION OF FREIGHT IN PAN AMERICAN TRAFFIC.

The object of the tentative plan is to facilitate the certification of freight in South American ports, when destined to United States ports, and vice versa, and to aid consular officers and medical officers of the various Governments concerned, when assigned to such duty, in deciding upon the treatment to be accorded to the various classes of freight.

1. Pan American quarantine conference.-A conference shall be held annually at some convenient point between representatives of the quarantine authorities of each country concerned, for the purpose of formulating and progressively carrying out the following expedients in each port tending to the economical handling of freight:

A. Reciprocal inspections of quarantine facilities, measures, and methods in the various Pan American ports.

B. Constant surveys of the rat population in the various ports, by means of organized rat collection and bacteriological examination. C. Progressive rat-proof construction and remodeling of wharf property, warehouses, and other buildings.

D. Certification of freight by selection as to character, methods of packing, sanitary conditions of locality where stored prior to shipment.

TENTATIVE DETAILS OF PROPOSED FREIGHT CERTIFICATION.

Certification of freight from localities infected or suspected of being infected with plague.-1. In the case of vessels which have been rendered rat free by fumigation just prior to receiving cargo, the following cargo may be loaded without treatment:

a. Filled, rat-proof containers, such as kegs, barrels, cans, boxes,

or cases.

b. Moist, or green dry, hides, if the containers are open to inspection to determine freedom from rats.

c. Upholstered furniture, general personal effects, secondhand articles, curios, feathers, mattresses, matting, wool and furs, crated bamboo ware, plants, bulbs and seeds, rope, cuttings, bones, tendons, bone meal, hides, horns, hoofs, hair or bristles, if it can be determined that the articles and their containers are rat free.