TABLE 3.-Summary of data in Table 2.

1. CITIES WITH LESS THAN 100.000 POPULATION HAVING TRAINING SCHOOLS.

II. CITIES OF 100,000 POPULATION OR OVER HAVING TRAINING SCHOOLS.

III. CITIES OF 100,000 POPULATION OR OVER WITHOUT TRAINING SCHOOLS.

PER CAPITA COST OF TRAINING SCHOOLS IN CERTAIN CITIES.

There is not sufficient material available at present to make a fair comparison of the cost of training schools. The figures given in the table are sufficiently suggestive to be worthy of consideration, however. Thus, in the eight schools shown, there is a range of nearly 400 per cent in the total per capita cost, but one can not be sure of the range of items included. In Baltimore, for example, the charge for fuel and janitor service for the training schools is made against the elementary schools occupying the same buildings. Then, too, the divisor used in determining the per capita cost makes much difference in the various cities. In some schools the proportion of graduates to enrollment is much smaller than it is in others. To give results that would be of real value in assisting school authorities to determine the relative cost in a number of cities, it would be well to show the per capita cost on the basis of graduates, as well as of total membership. Thus the amount paid by the city of Cincinnati for the work of the college of teachers, $10,500, if divided by the membership (243) would give a cost of $43, while a division on the number of graduates (33) would give $300.

There is special reason for working out formulæ for reporting the cost of small schools. The returns from some of the schools show very little charged against the training school except the salary of a principal, while a more exact accounting, such as was given in the 1911 report of the East Saginaw training school, shows an expenditure of $3,525 for 10 students.

TARLE 7.—Applicants, enrollment, and graduates of training schools.

1 The numbers in applicant's column for Baltimore denote admissions for graduation in year given.