492 DR JOHN MURRAY AND MR ROBERT IRVINE ON THE 



Another cause of loss of lime in the mud-waters may be the decomposition of 

 alkaline or earthy sulphides by bicarbonate of lime, the result being two molecules of 

 carbonate of lime in place of one, which will with difficulty be held in solution even in 

 sea- water, and will most probably be slowly deposited in the mud. In Table II., 4th 

 portion of filtrate, the carbonic acid (C0 2 ) was hardly above that of normal carbonate. 

 If we add the loss of carbonic acid to the salinity found, and also the carbonate of lime 

 precipitated, we find the salinity of the original water to have been 3 '405 4, which would 

 give a chlorine value equal to 18*859 grms. per kilo. — not far from 18*901, as found. 

 All these reactions or changes are, however, very complex and not easily disentangled. 



During the past year or two samples of sea-water have been collected for us from 

 various parts of the ocean.* The density, the chlorine, and the alkalinity have been 

 determined in each of these, and the values of D and D A have been calculated. The 

 results are exhibited in the following table, where 



Column a gives the locality from which the sample came. 



Column b, the density at zero temperature S . 



Column c, the density at 17°*5 C. i7- 5 S 17 . 5 . 



Column d, the chlorine % in grammes per kilogramme. 



Column e, the carbonic acid in milligrammes per kilogramme. 



Column f gives the D value ( -j) which was found from the density at 0° C. by the 



„ , oS - 10000 _ 



formula = D . 



X 



Column g gives the D A values, which is the density of sea-water minus 1000, divided 

 by the carbonic acid expressed in milligrammes per kilogramme, not per litre. 



In several instances the densities were determined at the temperature of 0° C, or that 

 of melting ice, and also at 17° '5 C. It was found that for ordinary sea- waters at any rate, 

 these two density values had a constant ratio, and the remainder of the waters were taken 

 at 17°*5 C, as being more convenient, and reduced by calculation to 0° C. It was found 



that 17 - 5Sl7 - 5 ~ 100 ° = 1*3800 (which agrees with Dittmar's results of D at 17°*5 C), 



when the D at 0° C is 1*4566. The formula applicable for reduction, after subtracting 



1000 from the densities, is 17 - sSl7 / 8 * * A 4566 = S . 



The mean D value at 0° C. of twenty-eight waters in the foregoing table, after strik- 

 ing out the abnormal waters (Nos. 2, 19, 20, 23, 24, 25, and 27) is 1*4566 (1*3800 at 

 17°*5 C.) ; which is probably not far from the mean value for the whole ocean. 



Dr Gibson from an examination of 122 waters from the Moray and Pentland Firths, 

 North Sea, and Arctic Ocean, obtained a mean D value of 1*4563, the maximum being 



* We are especially indebted to Dr W. S. Bruce, surgeon of the Antarctic whaler " Balaena," who brought us samples 

 in the spring of 1893, and to Captains Thomas S. Knox and George Reid of the Anchor Line for samples from the 

 Mediterranean, Red .Sen, Indian Ocean, and Atlantic. Mr Murray had still in his possession several samples collected 

 by the " Challenger." 



