ENGINEERING CONSTRUCTION AND RAINFALL. LXI. 



formed the essence of an empirical formula. Mr. Simpson's 

 quotation of Ganguillet and Kutter as giving many examples of 

 rivers with a velocity of 13 feet per second, and stating the 

 assumption on the part of the author of 6 feet per second as not 

 being in accord with facts, were due to misapprehension on Mr. 

 Simpson's part. As regards this particular subject, he was 

 quite aware of the existence of rivers with a velocity of 13 or 

 more feet per second, but in no part of his paper could he be 

 accused of advocating the application of a formula in such a case. 

 It was, as was pointed out, where a formula was applicable, and the 

 velocity of the water was unknown, as in the case of a creek 

 with intermittent flow, that he considered it better and 

 safer practice to assume a low velocity, for the reasons given. 

 As regards the charge of mathematical juggling preferred by Mr. 

 Simpson against the author, the latter was quite content to 

 leave the matter in the hands of those qualified to give an opinion 

 on this point. He agreed with Mr. Davis, and, indeed, so stated 

 in the paper, that reliable flood-marks were preferable to any 

 formulae. He could not agree as to the hard and fast 

 division of catchment areas into two classes, as proposed by Mr. 

 Davis j and as an old railway engineer could assure Mr, Davis 

 that such a course would be impracticable, however suitable it 

 might prove in the design of storm water channels. As regards 

 the typical case put forward by him, and referred to by 

 Mr. Davis, the latter was in error as to his meaning. 

 Q he illustration was given to point out the absurdity of grading 

 the rainfall : he had never intended that provision should be 

 made for a 6-inch rainfall over all the catchments between the 

 terminal points, but only such of these as, through the extent of 

 their areas and other circumstances, would render them liable to 

 such a rainfall. Mr. Davis's table of storm water channels, 

 which he produced to prove his (Mr. Haycroft's) assumption of a 

 co-efficient of run-off equal to unity, in the case of town 

 drainage, as being wide of the fact, was not pertinent, 

 as he would never think of proposing such a value 

 for areas of the extent given in the table. An analysis 



