flow of 224 gallons per minute, a maximum of 287 gallons, and a mean of 254 gallons.

Gaugings were made at Des Moines, Iowa, from June 30 to July 16, 1895, by J. A. Moore and W. J. Thomas, class of '95, Iowa Agricultural College (see Plate I). The sewerage system at the outlet of which the gaugings were taken comprised: on the west side 235,000 feet of sewers, contributary population 19,400, 15 hydraulic elevators; on the east side 29,000 feet of sewers, contributary population 8100, 3 hydraulic elevators.

These were combined sewers. Rain fell on two Sundays only, and is indicated by the unusual height of the curve. Water-meters were used on the services; water was supplied to 33,700 persons, but the amount consumed by each was not ascertained, the average consumption for the city being taken. The diagram for the west side shows noon-hour stops of factories. The high-water curves for July 10, 11, 12, and 13 were caused by the water company flushing dead-ends outside of the limits of the sewers gauged. On the 12th the large flow in the west-side sewer was probably caused by a part of this flushing-water reaching it.

The maximum dry-weather rate of flow on the west side was at 10.15 A. M. Friday, July 12-175.3 gallons per capita. The maximum dry-weather rate of flow on the east side was at 6.30 P.M. Tuesday, July 2-142 gallons per capita.

The minimum dry-weather rate of flow on the west side occurred at 4 A. M. Saturday, July 6-23.2 gallons per capita.

The minimum dry-weather rate of flow on the east side occurred at 4.30 A. M. Friday July 5-22.5 gallons per capita. The average dry-weather rate of flow on the west side was 66 gallons per capita.

The average dry-weather rate of flow on the east side was 74 gallons per capita.

Table No. 8 gives the water pumped and the sewage dis

charged during the seven days when the measurements taken were apparently reliable. The first column gives the total amount of water pumped; the second, the total sewage flow; the third, 81.6% of the first column, that being the proportion between the number of water-taps and that of the sewer-connections. The close correspondence between the two last columns shows what an excellent index the water-consumption furnishes, in this town at least, of the total house-sewage to be expected. July 12 was the only date when the maximum flow of the west side exceeded 125 gallons per capita, and then for two hours only. The average for this side was 66 gallons. Disregarding the maximum of the 12th instant, which was due to hydrant-flushing, we have a maximum for this side 89% greater than the average; and for the east side the maximum was 92% above the average. The record, however, covers two holidays out of the seven, making the average unusually low; also the general average for that time of year would ordinarily be lower than that for an entire year. It is thought that these include all the published records of gaugings of sewage flow which have been made in this country. They seem to point uniformly to the conclusions already stated that the winter flow of sewage is greater than the summer; that the maximum and minimum flow do not ordinarily vary from the yearly average more than 75%, but frequently do by 50%; that the house-sewage per capita very nearly equals the water-consumption where the taps and sewerconnections are equal in number.

The engineer must select and use with a great deal of judgment all the data obtainable in fixing upon the quantities which the sewer should be designed to carry. The method of making the calculations will be explained more at length in Chapter VII.

ART. 16. AMOUNT OF STORM-WATER.

The amount of storm-water reaching a given sewer depends upon the rate of rainfall, the time during which this rate is continued, the proportion of the rainfall which flows off, and the time taken by a raindrop after falling to reach the point under consideration. This last depends upon the shape, extent, and nature of the surface over which, and the length and grade of the sewer through which, it must flow.

ART. 17. RATES OF RAINFALL.

It is apparent that the rate at which the water reaches the sewer depends to a greater or less degree on the rates of rainfall from minute to minute, and not upon the amount falling in a day or even in an hour. Records giving rainfalls for these latter units of time are, therefore, valueless to the sewerage engineer. It is only within recent years that gauges have been used which automatically register the rate of rainfall at each moment of a storm. But so great a necessity for such records has been felt that the use of self-registering raingauges is becoming more and more general, and in most of the large cities continuous record of the rates of rainfall is obtained either by a city department or by the United States Weather Bureau.

Since the maximum amount of water to be removed determines the size of the sewer, we are concerned only with the maximum rates or those near the maximum. Rates of heavy rainfalls for various cities of the United States are

given in Table No. 9. Where possible several high rates. during the same or consecutive years are given for each locality, but no attempt has been made to give a record of all severe storms for any one place or year. An examination of rainfall data covering many years shows that in New England a rate of 3.6 inches an hour continuing for 5 minutes may be expected every year or two, a rate of 2 inches continuing for 20 minutes, a 1.5-inch rate continuing for 30 minutes, and I inch in 60 minutes. In New York State the rate may be about 20% and in Pennsylvania about 30% higher. In Baltimore and Washington we may expect a 5-inch rate for 5 minutes, a 4-inch rate for 10 minutes, a 2.7-inch rate for 20 minutes, a 2-inch rate for 30 minutes, and 1.4 inches in 60 minutes. In New Orleans a 5.5-inch rate for 5 minutes, a 4.5-inch rate for 10 minutes, a 3-inch rate for 20 minutes, a 2.5-inch rate for 30 minutes, a 2-inch rate for 60 minutes or even more may be expected. In the central States a rate of 3.7 inches for 10 minutes, 2.8 inches. for 20 minutes, 2.3 inches for 30 minutes, and 1.7 inches in 60 minutes may be expected. Further data, however, may require a change in any of these values.

Prof. Talbot gives as a formula of maximum rates of rain

fall in the eastern part of the country, r =

agrees quite closely with Plate IV S. Atlantic States up to 30 minutes, but gives too small values for longer periods.

The records seem to show, where any information on the subject is given, that the maximum intensity usually lasts but a few minutes, seldom more than ten; that it sometimes occurs at the beginning of a storm, but in a great majority of instances occurs at the middle or end of it, quite a number stopping 10 to 20 minutes after the maximum rate is attained. As to the area simultaneously covered by the maximum rates of fall, almost no data are available.