be planned for the entire area draining to any sewer for which calculation is made, tentative locations for both being made where street plans have not yet been laid out. Profiles are prepared of the surfaces of all streets, and the sewer-inverts are located on them in pencil, since the final calculation will probably result in minor changes in the sewer grades. The location of the inlet connections are indicated on the profiles. The rainfall records of the place in question are consulted to determine the rates of precipitation to be used in the calculation, and decision made as to what class of storms to provide for the maximum ever probable, the maximum omitting those not occurring oftener than once in ten years on an average, or storms of annual occurrence only. (Provision for the second class is believed to be advisable in most cases.) If such records are not available, use may be made of the rainfall curves in Fig. 7 or Fig. 22, or preferably of similar curves prepared from data for the region of the place in question secured from longtime records of the U. S. Weather Bureau. An imperviousness factor for each part of the area draining to each inlet is then decided on. In making this decision, the condition during the following thirty to fifty years should be considered. In general the tendency is for the imperviousness to increase, due to the changing character of street and sidewalk construction and the increasing number of houses with their impervious roofs. A plan commonly followed is to select a block fully developed as a business block, others as dense residence, medium residence and suburban blocks, and determine the actual conditions in such blocks. Then forecast which of these conditions each of the areas will attain within the time assumed. The safest course is to assume every street surface (both roadway and sidewalk) as wholly impervious, also all roofs. In business districts, the yards and courts should probably be given factors of .60 to .80. In residence districts, for lawns and gardens the factor would probably lie between 25 and 75, depending upon the porosity of the soil, assuming a rainfall of at least an hour preceding the maximum precipitation. The imperviousness factor for a given area may be calcu lated thus: Let 1=the average length of a city block; b= f= d= พะ breadth of a city block; number of front feet to a building lot; depth of a building lot; width of street percentage of imperviousness of yards, courts, etc., expressed as a decimal; percentage of imperviousness of the entire area, expressed as a decimal. 1= Then As an example, let l=450, b=250, f=50, d=125, a=1200 square feet, w=66, i= .60; then I = .777, or say .78. When most of the above factors must be estimated by judgment only, as for areas not yet opened up or fully developed, it may be as well to estimate I at once. A map is used on which are indicated the locations of all inlets and sewers, and on each area its imperviousness factor I and its area a. The run-off from each area to its inlet is then calculated as in Art. 11. Since we are assuming future conditions, all of the blocks in a given district will be similar except as to dimensions and grades of the bounding streets; and if the street lay-out is at all regular, the blocks may be arranged in a few groups so nearly alike in these respects that one calculation will suffice for each group. The areas draining to inlets do not ordinarily differ in their run-off times nearly so much as do the areas between successive run-off contours of a single block; consequently there will be much less error in assuming that the precipitation rate remains constant during the run-off time assumed for a district including a group of blocks. The method employed in calculating sewer capacity for a number of blocks may therefore be simpler. The following calculation illustrates a method of making and tabulating a typical problem, reference being made to the map, Fig. 23. AI is in each case the sum of all the preceding al's; is the greatest distance traversed by the run-off in crossing t is the time occupied by the run-off in travelling the distance l; q=alr; S is the slope of the sewer removing the run-off from the point in L its length to the point next considered (usually the next inlet con- T is the time occupied by the run-off in flowing from the extreme R is the rate of rainfall for the time T; Qis the total amount of run-off from all drainage-areas above=AIR. The quantity q(=aIr) as well as Q should be calculated for each sub-area, and if the Q for any stretch of sewer is at any place less than the q immediately tributary to the same, the latter should determine the size. Fig. 22 will be found convenient for determining a, and also R when the rates of rainfall of the place in question can be represented by any of the curves there given, these being for rains at ten-year intervals. To find a in acres from the diagram, use one dimension (in feet) of the area (or of an equivalent rectangle if it is not rectangular) as an ordinate and find the corresponding abscissa of the acre-curve in the diagram; divide this into the other dimension of the area and the quotient will be a in acres. By the table, the run-off from the undeveloped territory is placed at 60.0 cubic feet per second, which is carried by a 42-inch sewer on a .6 per cent grade for 360 feet, where it receives still more sewage; the maximum amount to be received there, both over the surface and through the sewer, being 62.4 cubic feet, although q for the block No. 1 alone is 7.6 cubic feet. But Q is not equal to 60.0+7.6, because the latter quantity was due to a rainfall of 3.9 minutes' duration, or rather to the maximum rate for that time, during which only |