AGE OF EARTH AND SALTNESS OF SEA 191 



into it each year is a much more doubtful quantity. Into all the 

 sources of error it is not possible to enter here; but the following 

 short summary will show that these are very great. 



In the first place we must remember that sodium is a minor con- 

 stituent in river waters. It is also, as every chemist knows, the most 

 difficult of all to determine. Anyone who is acquainted with the 

 many possibilities of error implicit in the process of boiling down 

 considerable quantities of river water, removing all the other con- 

 stituents in order to find in the end possibly two or three parts per 

 million of sodium, will know that very little reliance should be placed 

 on this part of an ordinary water analysis. In order to obtain 

 results of any value for this constituent, great care and special pre- 

 cautions must be taken. So far is this from being the case in the 

 average analysis, especially the older ones from which Sir John 

 Murray's tables are constructed, that all the alkali metals are com- 

 monly determined together and stated as sodium and potassium. 



Once more, Professor Joly makes an entirely inadequate allowance 

 for cyclic salt. Sea salt, especially in stormy weather, continually 

 passes into the air, is brought down by the rain, and is deposited 

 in other ways. This is washed into the rivers and thus reaches 

 the sea once more. Another form of cycHc salt is that known as 

 fossil sea salt. This was deposited with the other constituents in 

 the sedimentaries and is once more set free by erosion. As none of 

 the ordinary rocks (with the exception of salt beds) contain any 

 appreciable proportion of chlorine, it is highly probable that the 

 greater part of the chlorine in river waters is cyclic in some form or 

 other. This particular aspect has already been the subject of some 

 controversy, and Professor Joly has admitted that he has probably 

 underestimated the amount of cyclic salt; but he meets the difficulty 

 by stating that there is a considerable preponderance of sodium in 

 the river waters above the chlorine, and that, even assuming that all 

 the chlorine is cyclic, the estimate of geologic time is not thereby 

 raised to more than one hundred and forty millions of years. ^ 



But, by this admission, the whole calculation is made to rest to a 

 much greater extent on the minute accuracy of sodium analyses, 

 which, as we have already seen, does not exist. To show what differ- 



I See discussion between Professor Joly and Mr. Acroyd, Chemical News (1901). 



