ENGINEERING CONSTRUCTION AND RAINFALL. XLV. 



the water reaching the outlet from the farthest point of the 

 catchment, whilst it still rained. 



In view of the above considerations, it is evident that not one 

 of the formulae proposed is entirely general^; are they then to be 

 discarded, and, if so, what course should an engineer pursue when 

 designing a waterway through a bank ? Before answering this, 

 the following extracts from standard works and leading en- 

 gineers as to practice in other parts of the world, may be noticed. 



In Patton's Civil Engineering, under the head of " Water 

 Reaching Streams and Sewers," it is stated : — 



" The run off depends on such conditions, that the variations of 

 formulae as well as the difficulties of applying them to small city areas and 

 large country areas alike, make the application unsatisfactory ; formulae, 

 now applying approximately well for city areas, do not apply to country 

 areas, where the storm discharges are carried off by creeks and rivers ; the 

 best formulae now used seemed to be based on variable areas, variable slopes 

 and variable rainfalls, the powers, roots, and constants, used in each, giving 

 it its special merit." "Even with the best 4 formulae for run off, 3 of 

 them give curves for areas under 5 acres, shewing more run off than rainfall, 

 and yet Biirkli Ziegler and McMath are more generally used, because of 

 better agreement with observed run off from areas, say above 50 or 60 acres ; 

 none of these formulae, and still worse, none of the various flood discharge 

 formulae, are satisfactory in very large country areas." 



Professor Johnson, of Washington University, from whom the 

 writer sought information, states, as regards design of water- 

 ways : — 



" There is no fixed practice, and any engineer would find it hard to give 

 any fixed rules for his own practice. Some roads always run out the 

 watershed line, and find the drainage area of the stream or draw crossed, and 

 then assume a maximum rate of rainfall, and use some formula for getting 

 the maximum rate of run off from the area ; character of surface and rainfall 

 rate assumed, but in the end it is little better than a guess. We have a 

 tradition here amongst our Eailway Engineers, that it is only a question of 

 time before any culvert will ' go out ' by a flood, so you see we have no ' safe 

 rule ' to go by. Make them as big and as permanent as you can afford to do 

 is the more common American practice, and if they go out ' charge it up to 

 Providence.' 



Mr. G. H. Pegram of the Union PaciBc System, states : — 

 " The general practice has been for the locating engineer to size up the 

 situation at a glance, as he goes over the road, and put down a ' 2 foot box ' 



