tive importance of salt beef is 11 in the South Central division, while in the North Atlantic division it is 242, or 22 times its importance in the South Central division.

The importance of salt hog products is 361 in the North Atlantic division, 897 in the South Atlantic division, 350 in the North Central division, 960 in the South Central division, and 359 in the Western division. The importance of sugar does not show a great difference; it is 493 in the North Atlantic division, 474 in the South Atlantic division, 455 in the North Central division, 536 in the South Central division, and 470 in the Western division.

By the use of this table a series of index numbers, or relative prices, has been computed in which the principal articles of food have been given the weight or importance shown in the preceding table, thus showing the course of retail prices from year to year when the various articles are weighted according to average family consumption.

Certain groupings of the articles for which retail prices were secured are necessary in order to bring them within the classification reported in the family budgets.

Fresh beef is represented by the two articles-one, roasts, and the other, steaks; salt hog products by the three articles-salt bacon, dry or pickled pork, and salt ham; other meat by the two articlesmutton and veal; poultry by the single article-chickens; fish by fresh fish and salt fish; flour and meal by wheat flour and corn meal. However, in obtaining an average index number to be used in weighting by relative importance, wheat flour is given twice the importance of corn meal. It would be necessary to give wheat flour even a greater weight as compared with corn meal but for the fact that bread is considered separately from flour and meal. Fruit is represented by two articles only-evaporated apples and prunes-as it was impossible to secure complete series of prices of other descriptions of fruits. Vinegar, pickles, and condiments are represented by the single article vinegar.

This list includes all articles for which retail prices were secured except beans. The groupings are as follows:

The figures showing the relative importance of the various articles of food are based on the family consumption for 1901. There is no doubt that the relative importance of articles varied slightly from year to year, but no figures are available covering the 17 years, therefore the relative consumption of the various articles of food is, in this study, considered the same for each of the years.

The method of computing the series of relative prices weighted according to family consumption, for each of the geographical divisions and for the United States considered as a whole, can best be explained by illustrating the process.

The following table is for the United States for the year 1906:

TABLE ILLUSTRATING THE METHOD OF COMPUTING THE RELATIVE RETAIL PRICE OF FOOD IN THE UNITED STATES IN 1906. THE VARIOUS ARTICLES BEING WEIGHTED ACCORDING TO AVERAGE CONSUMPTION IN 2,567 WORKINGMEN'S FAMILIES.

The sum of the numbers indicating the relative importance of the various kinds of meat is 2,845, the relative importance in family consumption of all meat. The sum of the results which were obtained. by multiplying the relative price of each of the kinds of meat by the number indicating its relative importance is 350,382.5. Therefore to obtain the relative price of the group, meat, for the United States. in 1906, the sum of the results is divided by 2,845, the relative importance of the meat group. The quotient, 123.2, shows the relative price of the meat group in 1906 when weighted according to family consumption.

Each of the remaining 17 articles or classes of articles is taken up in the same way as were the several kinds of meat, the relative price. for the United States in 1906, as shown in Table II, being multiplied by the relative importance of the article, as shown in the table on page 198. For the group of 17 articles comprised in "total, except meat," the relative price, 112.4, was found in the same way as was the relative price for all meat, the sum of the results, 669,261.0, being divided by 5,954, the sum of the numbers indicating relative importance. One other group, designated "other food," remains to be considered. As the various descriptions of meat most generally used and also a class, "other meat," were included in the total meat, it is presumed that all meat was included in that total and that the food which in the family budgets was classed as "other food" must be more nearly related to the various articles considered in the group "total food, except meat," and therefore "other food" was given the same relative price as "total food, except meat."

The grand total of the numbers indicating relative importance was 9,423, and the sum of the results obtained by multiplying the relative price of each of the various articles considered by its relative importance was 1,089,781.1. The sum of the results divided by 9,423 is 115.7, the relative price of food for the United States in 1906, when the various articles are weighted according to the reported consumption in 2,567 families. The class "other vegetables" has a relative importance of 577, but it was impossible to collect retail prices of articles to represent properly this class for the sixteen years; hence it was not possible to include that class in this table, and therefore the grand total of relative importance, 9,423, lacks 577 of making the 10,000 total importance of food indicated in the table on page 198. That is, 94.23 per cent of the total food consumption per family is represented in this table.

It is interesting to see, from the table on page 179, that the relative price for 1906, weighted according to family consumption and including 94.23 per cent of the total food consumption, differed but little from the simple average of the 30 articles of food shown on the same page, the weighted price being 115.7 and the simple average 116.2.

It should not be understood that the 30 articles for which retail prices were secured, and which are enumerated on page 204, comprise 94.23 per cent of the food consumed, but these articles represent groups or classes which do, according to the 2,567 family budgets, comprise 94.23 per cent of the food consumed.

The weighted relative price for the United States for each of the years from 1890 to 1905 was found in the same way as has been explained for 1906.

The same method was employed in computing the weighted relative prices for each of the five geographical divisions as for the United States. The relative importance of the various articles, or classes of articles, in the different geographical divisions is shown in the table on page 198.

The following table shows the average price per pound, dozen, quart, etc., in 1901, of various articles of food. One column shows the average price paid by the 2,567 workingmen's families investigated in the United States, as reported by the housewife or other member of the family—that is, the average amount paid per family for each article was divided by the average quantity of the article consumed. Another column shows the average price computed from the reports obtained from the retail merchants. The average price shown for fresh beef in the last column, for instance, is the simple average of the prices quoted on the 614 schedules for fresh beef; the average for salt beef is the simple average of the prices quoted on the 121 schedules for that article, etc.

AVERAGE PRICE PER POUND, DOZEN, QUART, ETC., OF CERTAIN PRINCIPAL ARTICLES OF FOOD, 1901.

The average price per pound for fresh beef reported paid by the 2,567 families was 14.31 cents; the simple average price from the 614 schedules for fresh beef secured from the retail merchants was 14.70 cents. The average price per pound for salt beef reported in the family budgets was 10.82 cents; the average from the 121 schedules for salt beef secured from the retail merchants was 9.31 cents. In the family budget salt beef included dried beef, while the prices from retail merchants are for corned beef. The prices for the other articles can be compared by referring to the table.

It will be noticed that for all articles except four-salt beef, cheese, sugar, and flour and meal-the average price computed from the family budgets is less than the reports of retail merchants. In explanation of this difference, two suggestions may be offered. The first is, that when the price of any article of food advances there is a tendency on the part of the consumers to use some substitute for that article. Therefore, if an article is low for six months of the year and high for the remaining six months the greater quantity consumed by the family may have been purchased during the months when the price was low, thus producing a lower average for the year than would the simple average of twelve monthly quotations. A second suggestion is that prices secured from retail merchants were usually for some standard grade or brand of article, while in some families investigated it is no doubt true that articles of a lower grade were purchased at times.

The table which shows for each of the 5 geographical divisions and for the United States the relative price of food for each year from 1890 to 1906 is presented on page 179. The results were computed by two methods. By one method the figures obtained are the simple average of the relative prices of the 30 articles, and by the other method the relative prices of the various articles were weighted according to their average consumption in the workingmen's families investigated in the States included in the geographical division.

EXPLANATION OF SCOPE AND METHOD.

The present report gives the results of an investigation of retail prices in 69 of the principal cities of the United States for the year 1906 and also summaries covering the seventeen-year period from 1890 to 1906. Previous investigations of retail prices have been made and published by the Bureau of Labor, as follows: Eighteenth Annual Report, covering 1890 to 1903; Bulletin No. 59 (July, 1905), covering 1904; Bulletin No. 65 (July, 1906), covering 1905.

The present investigation includes 6,176 schedules or statements of prices secured from 993 firms in the principal industrial localities in 39 States, including the District of Columbia. In the Eighteenth