LVI. DISCUSSION. 



formula would decrease from 2 to J. The author, in making 

 the calculations shown in his table, had made R constant 

 throughout. He wished it to be clearly understood, therefore, 

 that he rejected the whole of these calculations as erroneous. 



Prof. Kernot regarded the paper as a very complete and 

 valuable compendium of what had been so far settled or con- 

 jectured on a very complex subject. With regard to his own 

 formula, it was at best rough, and to be used only under limita- 

 tions, yet it very fairly corresponJed to the most satisfactory 

 practice coming under his notice. Melting snow was a factor 

 which, under certain circumstances, might be of great importance ; 

 in the case of warm rain in spring falling on snow it might be 

 allowed for by an increase of 50% in the constant of his formula. 

 He thoroughly agreed with the author's advocacy of self- 

 registering rain-gauges. Very little had been said with regard 

 to storage, but this in many cases appeared to give considerable 

 relief, and permitted waterways to be reduced in size. As to the 

 velocity of discharge through culverts, he had often gauged it in 

 actual cases, and found it rarely to exceed 8 feet per second. 

 Such velocities as 18 feet per second he considered utterly inad- 

 missable, except in the case where everything in the vicinity was 

 solid rock or masonry. The efficiency of culverts might be raised 

 by making their inlets and outlets approximate in form to a vena 

 contracta. He agreed that the co-efficient of run-off should be 

 unity when the catchment consists of roofs, pavements, and hard 

 metalled roads. 



Mr. J. Davis thought that the author was to be compli- 

 mented for his industry in collecting so extensive a list of 

 formulas. The paper was, however, disappointing, because, while 

 the author fully appreciated and weighed the difficulties of the 

 case, he practically left it in much the same position in which he 

 found it. In considering the provision to be made for the dis- 

 charge of water under railways, streams should be divided 

 into two classes — viz., (A) those flowing from catchments over 200 

 acres, and, therefore, of comparatively large volume ; and (B) 

 those flowing from areas less than 200 acres. Usually in the 



