Many parts of telephone plant, such as central office buildings and equipment, conduits, underground and aerial cable at the time of installation must have the capacity to handle not only the immediate demand for telephone service, but also to take care of growth for a number of years to come. In order to engineer such items of telephone plant economically it is necessary to know in advance as accurately as possible what the demand for telephone service will be five, ten, or twenty years in the future. Forecasts of the future market are very necessary for plant engineering, operating plans, rate treatment, and other purposes, in multi-central office cities. In such cities detailed estimates are made of the market some twenty years ahead and of its telephone development under stated rate conditions. Such estimates are called commercial surveys, and they involve a study of the various factors which, in the course of events, will be likely to control the industrial, commercial and residential development of the city concerned. In the course of such a survey, a rental classification of all families is obtained and at the same time a record is made of existing telephone service in each rental class. The rent data of this article have been gathered in representative large cities throughout the country and the results as here set forth are being used together with many other kinds of data to guide the engineering of future additions to the plant of the Bell Telephone System. In general the income of a family is an index of the market it creates for various commodities including telephone service. Rental values may also be considered as such an index and the present study seeks to correlate rents with incomes. Rents can be readily recorded and classified, whereas it is not feasible to determine the money incomes of large numbers of families. While it may be ideally possible, by a study of rent data, to compare the inherent markets for telephone service and also the strength of the telephone habit in various cities, there are many practical limitations to such a procedure. Comparison of the residence market for telephone service in different cities, as determined by rent values, is made difficult by the fact that the variation between cities in rentals paid for substantially similar dwellings is considerably greater than the variation in prices for food or clothing. Further, there is considerable variation in rent levels even in different sections of any one city. Attempts to compare rent distributions by application of the usual statistical measures of dispersion and skewness have proved unsatisfactory. However, a method of charting has been found by which rent distributions may be readily compared with one another and an index of spread or dispersion determined. It has been found that cumulative curves of rent distribution may be plotted on logarithmic probability paper to yield straight lines for a large number of cities. These are called logarithmic skew distributions. Although it has not been found possible to assign any special significance to the particular value of the index of rent dispersion in any city, this index appears to remain practically constant for that city regardless of changes in the level of prices. In the appendix the mathematical features of the logarithmic skew curve are discussed. — Editor.