The purpose of this paper is to examine the questions raised by John W. Alexander (1954), by analyzing 97 cities in Japan with the index of surplus workers as shown below:<br> _??_<br> where S refers to the number of surplus workers (basic activities), C to each city, N to the nation, i to the number of employed persons of each industry, and t to what demands each industry. In this paper employed persons working outside their resident city are aho considered surplus workers because they bring their wages into the city.<br> In the above formula John M. Mattila and Wilbur R. Thompson (1955) used total employment for t. If we use what demands directly each industry in place of t however, we can obtain a more accurate measure of basic activities. For that purpose we first convert the 1970 Input-Output table in terms of producers' prices into the I-0 table in terms of employment. Then assuming that the individual demand sectors of the table, except the three sectors of primary industry, increase in stock, and export, are what demands each industry, we calculate the number of employed persons each city needs, Di(Ct•Ni/Nt), with the ea uation shown below.<br> _??_<br> where j refers to each of the individual demand sectors, and Xij to employment in industry of the sector j. The letter t refers to employment in each industry as to each of the individual intermediate demand sectors, to population as to private consumption-expenditures sector, to the number of office workers of non-primary industry excluding government as to both consumption-expenditures outside household and gross domestic fixed-capital-for-mation sector, and to the number of office workers of government as to government con-sumption-expenditures sector.<br> Cities for this paper are selected through the following procedure;<br> We can use only the statistical data of administrative city unit in Japan. However, the administrative city is hardly in accordance with the urbanized area. In many cases the former extends far beyond the latter, and the disaccordance between the two areas varies from city to city. In order to get the criterion for selecting cities, we analyze the industrial structure in terms of employment of densely inhabited districts (DIDs). The DID is instituted as the sole statistical urban area in Japan since the 1970 Population CensLis. The primary industry ratio is under 5% in 80% of 565 DIDs within 579 adminbtr ttive cities.<br> The administrative cities are divided into two groups, those located within Tokyo or Osaka metropolitan region and the others. In the first roux. there are 32 cities whose primay-industry ratio is under 5 per cent, and all the 52 cities are selected for the analysis. For the cities in the second group, the criterion ratio is raised to 7%=for the cities with more than 300, 000 people, and to 9 per cent for the cities with 200, 000 to 299, 999 people to make the difference between the numbers of selected cities of the two groups as small as possible. Forty-five cities are selected from the second group in the above manner for the study, waking the total number of selected cities 97.<br> The outline of the results that were obtained by the analysis is as follows:<br> 1) As the population size of tie city increases, the N/B ratio tends to increase. When we divide the 97 cities into 44 independent cities that have less than 100 employed persons working outside the city per 1, 000 inhabitants and 53 dependent citis that have more than 100, we find that the above-mentioned trend is more distinct (Fig. 2). Generally, the N/B ratio of an independent city is higher than that of a dependent city of the same size, <br> 2) The N/B ratios of the independent cities show a discontinuity at the population of ap-proximately 150, 000.