192 H. S. SH ELTON 



ences small errors can make, it is interesting to note that more recently 

 another worker, M. Dubois, has made fresh calculations, using 

 only the more accurate sodium analyses, and has obtained as the 

 result four hundred millions of years.' I, for one, should not like 

 to state dogmatically that one or other of these results is accurate; 

 but certainly the calculation of Professor Dubois is the more reliable 

 of the two. It is more instructive to note the reason why such 

 diverse estimates are possible. This can be seen by anyone who 

 refers to the original paper and makes a few simple calculations. 



According to Sir John Murray's data, the percentage of sodium 

 in river waters compared with the total dissolved solid is 3.47, that 

 of chlorine 1.85.^ Professor Joly's calculation assumes 10 per 

 cent, of the chlorine to be cyclic. If we assume the amount of 

 sodium to be 2 . 47 per cent, (an amount of error not at all unlikely 

 to occur in these analyses, especially if calculated in a rough statis- 

 tical way), and the more probable quantity of 90 per cent, of the 

 chlorine to be cyclic, our estimate would be raised to two hundred 

 and fifty millions of years. If the sodium were but i . 5 per cent, 

 of the total solid, the estimate would be fifteen hundred and seventy 

 millions of years. As we have seen, M. Dubois' estimate was four 

 hundred millions of years, but he has reached such a point in the 

 calculation that a very small error would make a great difference in 

 the result. From such considerations as these we can understand 

 the uncertainty of any such estimate on present data. It is to be 

 hoped that, in the future, chemical analysts will pay special attention 

 to this problem of the proportion of sodium, and that, in this way, 

 more accurate data may be obtained. 



Nor must it be assumed that, however accurate may be the data, 

 this problem can be regarded as solved. There still remain a number 

 of theoretical objections. The method assumes approximate uni- 

 formity in this process of the conveying of sodium to the sea. It 

 also assumes that the sodium which reaches the sea never returns 

 to the sedimentaries. Neither of these assumptions, though prob- 

 able, can be regarded as established. Our knowledge of geological 

 chemistry is certainly not sufficient to enable us to say that none of the 



I Proc. Amsterdam Academy (1902). 

 ' 2 See Clarke, Data of Geochemistry, p. 88. 



