1906] 



IN CHEMICAL GEOLOGY. 31 



nitudes of his corrections are uncertain, and secondly, our knowl- 

 edge of the contribution made by rivers to the sea is most imperfect. 

 Joly has used Sir John Murray's estimate of the composition of river 

 water,^ and that rests upon insufficient data. Murray has averaged 

 together the analyses of nineteen rivers, which are not named, but 

 which were presumably, for the most part, European. Data are 

 lacking for the great African and Asiatic rivers, the Nile excepted, 

 and the other available material is incomplete. A river varies in 

 composition from time to time and from place to place ; so that 

 a single analysis of it may be misleading in the highest degree. 

 And yet many of the published analyses are of that character ; 

 that is, they represent single samples of water from river systems 

 in which the local and annual variations may be very large. Further- 

 more, Murray's average is mainly, if not wholly, of waters from the 

 temperate zone ; from which, in all probability, tropical and arctic 

 waters may differ considerably. I speak thus advisedly, for I have 

 compared, and reduced to uniform standards, more than a hundred 

 analyses of river waters, and have noted their wide variations. For 

 a very few streams the average annual composition is known, and 

 from such averages, weighted in accordance with the areas drained, 

 the final estimate must be made. The data are now being gathered ; 

 the nature of the work to be done is well understood, and in a few 

 years it may be possible to replace Murray's average with one of a 

 more definite character. Then, and not till then, can Joly's method 

 be profitably applied to the discussion of geological time. His results 

 may or may not be seriously modified, but the conclusions reached 

 will be more definite than any we can attain to now. At all events, 

 the present rate of chemical denudation is a measurable quantity, 

 and it will form an important statistical datum for use in the inves- 

 tigation of various problems. 



In conclusion I may be permitted to urge upon chemists and 

 geologists the importance of the statistical method in the investi- 

 gation of large geochemical problems. The method is evidently 

 applicable in many cases, and leads to conclusions of positive value. 

 We need, however, better material to work with than we have now, 



^Scottish Geogyaphical Magazine, 1887, p. 76. 



