450 



Part III — Seventh Annual Report 



tailed and discussed in my previous reports. For the sake of continuity, it 

 may not be out of place to give here a brief account of my plan of work, the 

 more so as a clear understanding of it is necessary for the purposes of the 

 following/liscussion of the results arrived at up till now. According to the 



S - W 



formula given above, viz., * t 4 t = D the c liloririe, or rather total halogen 



X 



calculated as chlorine in unit weight of any given sea water, bears a constant 

 relation to the excess of density of the sea water over that of distilled water 

 of the same temperature. Supposing sea water to be diluted with pure 

 water, while the percentage of chlorine would of course be lowered, this 

 relation would remain unchanged, and D would have the same numerical 

 value. 



Supposing, however, sea water to be diluted with river water con- 

 taining a sufficient quantity of saline marter, of different quantitative 

 composition from that of sea water, the relation between total halogen and 

 the excess of density of the water thus diluted over that of pure water at 

 the same temperature would be altered ; in other words, D would be 

 represented by a different number. Therefore, in order to determine 

 whether or not a given sample of sea water contains salts, or more 

 generally matter in solution of appreciably different quantitative composi- 

 tion from that to be found in normal sea water, all that is necessary is to 

 determine with a sufficient degree of accuracy the percentage of total 

 halogen, and the specific gravity of the sample in question. It is possible 

 that an alteration in the value of D should be quite appreciable, and 

 yet the alteration in the relative proportion of any one constituent be 

 almost if not quite inappreciable even to the most refined methods of 

 analysis, because the alteration in the value of D is the result of the 

 alteration in the relative proportions of all the constituents of the dis- 

 solved matter, and not merely of any one of thern. The addition to sea 

 water of any mixture of soluble matter actually derived from the drainage 

 of the land necessarily teuds to increase the value for D. 



It is evident that if the determination of the value for I) is to be of 

 any real value the determination of the specific gravity must be made 

 with a degree of accuracy corresponding to that of the determination 

 of the amount of chlorine, or total halogen calculated as chlorine. This 

 cannot be said, however, of the great majority of the specific gravity deter- 

 minations of sea water hitherto published. Hydrometers can indeed be 

 made very delicate, but it does not follow that the accuracy of their 

 indications is at all on a par with their delicacy. Another source of 

 uncertainty is that such hydrometer readings only become comparable 

 when corrected to a common temperature, that is to say, when allowance 

 is made for the alteration in the volume of sea water with change of tem- 

 perature. If sea water behaved in regard to change of temperature in the 

 same manuer as pure water this would be no real difficulty, as the 

 coefficient of expansion of pure water has been determined with a more 

 than sufficient degree of precision. 



Sea water, as is now well known, does not behave in this respect pre- 

 cisely as pure water, and the determination of the law of its change of 

 volume with change of temperature has been made the subject of investi- 

 gation by several authorities ; but unfortunately the results are not in 

 sufficiently close agreement, nor has it been proved that they are applicable 

 to sea waters affected by admixture with river water. 



For the reasons above stated I thought it better to adopt a method 

 not liable to the various sources of error which infect all hydrometer 

 determinations. "Without going into any details, for which I must refer 



