50 J. nOPKINSON HERTFOBDSHIRE EAINFAXL, 



the difference being IJ, ^, and ^ per cent, respectively; not nearly 

 80 closely in any decade with the ratio of annual rainfall ; and 

 that it does not correspond at all with the ratio of summer rainfall. 

 If the ratio of percolation be calculated on the mean proportions 

 to the rainfall in summer and winter at jSTash Mills and Lea Bridge, 

 that is 6 per cent, of the summer rainfall and 46 per cent, of the 

 winter rainfall, the ratios of percolation in each decade should be 

 as follows: — First decade, 96 per cent.; second, 95h per cent.; 

 third, 98 per cent; fourth, 109 per cent. ; and fifth, 101^ per cent. 

 The differences between these calculated ratios and the actual ratios 

 as given in the above table, for the second, fourth, and fifth decade, 

 are respectively ^, 1, and 1 per cent., thus showing that it makes 

 very little difference whether we estimate the percolation on the 

 proper proportion of summer and winter rainfall or on the winter 

 rainfall alone. 



The much higher ratio of j)ercolation than of rainfall to the 

 mean of the whole period in the first decade seems to be due to 

 excessive percolation during the first few years of the experiments 

 at Nash Mills. There must always be an undue amount of perco- 

 lation for some time after the natural soil has been disturbed. 

 The same thing is shown at the commencement of the experiments 

 at Lea Bridge, but here the normal percolation was reached much 

 sooner than at Nash Mills. At Rothamsted, where the soil was 

 not disturbed, the gauges being built up around it, there was no 

 excessive percolation at the commencement of the experiments. 

 The small ratio of percolation in the third decade, which is the 

 mean of the ratios at Nash Mills and Lea Bridge, is due to there 

 being a very small amount of percolation at Nash Mills during two 

 years of rather small rainfall. The character of the rainfall may 

 be the cause of this, but that could only be ascertained by an 

 examination of the daily fall. 



The period from which the mean ratio of percolation at Rotham- 

 sted is taken (the last 21 years) having a rainfall 5 per cent, above 

 the average for the half-centuiy, all the ratios during this period 

 are lower than they would have been if the experiments there 

 had been carried on for the half-century, for the mean percolation 

 would then have been less. For the last eight years the ratios are 

 to the mean of the percolation at Rothamsted only, and therefore 

 they are a little too low, as will doubtless appear when the results 

 of experiments at Nash Mills and Lea Bridge during these eight 

 years are made known. 



In calculating the areas of the different catchment-basins, and 

 the probable yield of water from the rain which falls on them, 

 I have assumed hitherto that the water-partings on the surface of 

 the country correspond with the water-partings underground ; that 

 is to say, that the rain which percolates in each catchment-basin 

 keeps within that basin after it arrives at the plane of saturation 

 in the Chalk ; but there are good grounds for inferring that such 

 is not always the case. Some of the rain which falls in the upper 

 part of the Colne basin almost cei'tainly is carried underground 



