E. J. Russell and E. H. Richards 311 



collection of samples at the gauges was done by E. Grey, who happily 

 is still continuing the same work. Warington and Miller differed widely 

 in personal characteristics but they were equally reliable analysts and 

 equally undeterred by the monotony of the routine involved in the 

 systematic work. Miller indeed made it his life-work, and only his 

 sudden and unexpected death in January, 1917, prevented the com- 

 pletion of an important monograph which he had in preparation on the 

 composition of rain and drainage waters. 



Miller's results show that the data of 1855 and 1856, which had been 

 wholly inadequate to give any support to Liebig's view, were probably 

 in excess of the actual facts; the ammoniacal nitrogen apparently 

 having been over estimated and given as 6 or 7 lb. per acre instead of 

 2-5 lb. to 3 lb. ; during Miller's period the sum of the ammoniacal and 

 nitric nitrogen was only 4 lb. per acre, — an amount still less adequate 

 to supply the needs of crops. 



Miller published his first paper in 1905^, but since then a further 

 ten years' results have accumulated which it is proposed to discuss here. 

 The work has now been modified: no useful agricultural purpose would 

 be served by continuing it in its original form, and its interest now lies 

 in its relationship to atmospheric pollution. For the Rothamsted rain 

 is collected in a part of the country which is fairly free from sources of 

 impurity, and it may be taken as typical of "pure" rain: it thus affords 

 a basis for estimating the amount of pollution. From this point of view 

 there is the possibility of a useful continuation of the work. 



The amount and distribution of the rainfall at Rothamsted. 



The rainfall records at Rothamsted extend continuously over a 

 period of 66 years, the results being 



The average for the year over the whole period is 28-3 inches. It will 

 be shown later, however (p. 324), that the rainfall tends to increase: 



» This Journal, 1905, 1, 280-303. 



21—2 



