REPOET ON PHYSICAL INVESTIGATIONS. 273 



surface water were collected. These have been subjected to the most 

 careful examination of which the resources of the Laboratory will 

 permit. The total halogen, calculated as chlorine, and the alkalinity 

 of each sample has been determined, as well as the density both by 

 Buchanan's hydrometer and the Sprengel tubes. The examina- 

 tion of these samples has shown that the water over the area under 

 consideration is normal Atlantic water throughout. The mean value 

 of " D " i. e. the ratio of the excess of density of sea-water at C. 

 over that of distilled water at the same temperature, to the amount of 

 chlorine in the sea- water, is 1*4553, with the limits r450 and 1*457, 

 a result in complete accordance with the value found by Dr. Gibson 

 for what he believed to be Atlantic water, contrasted with water from 

 other areas, as for example the Arctic Ocean, which gave a 

 markedly higher value. The close agreement of these determinations, 

 made under considerable disadvantages, affords strong proof of the 

 importance of a thorough examination by a chemical specialist of a 

 limited number of samples of water from the great oceans of the 

 globe. It seems possible to determine, once for all, the values of 

 D for each of the great oceanic basins, and from these to ascertain 

 at any time the source of supply of special currents in any particular 

 area. 



The chlorines of twenty-seven of the samples collected in June 

 last were determined before the publication of the previous paper, 

 but were not included in it, as it was then intended to make a series 

 of density determinations with the Sprengel tubes. It was, however, 

 found impossible to make these determinations, and a comparison 

 was therefore made between the Sprengel tube determinations of the 

 second series and the densities of the same samples as found by the 

 hydrometer, reduced to 15*56° C. as compared with pure water at 4 

 C, and then again reduced to 0° C. and referred to pure water at 

 0° C, by the help of Dittmar's tables. The results thus calculated 

 differed from the Sprengel-tube values by — 0*00013, giving a 

 correction almost identical with Dittmar's " hydrometer error." It 

 would seem, however, that the numerical agreement is accidental, 

 and these experiments have led to a more extended investiga- 

 tion at present in progress. Accepting the correction +0*000] 3 

 for the present, and applying it to the hydrometer determinations 

 of the June samples, the mean value of D is 1*4550 with the limits 

 1*453 and 1*457, a further confirmation of Dr. Gibson's results. 



So far as the present inquiry is concerned, it may therefore be 

 assumed that the water of the English Channel is Atlantic water 

 pure and simple, and the alkalinity determinations do not show any 

 variation in its " strength " by dilution with fresh water or other- 

 wise. The chief interest accordingly centres round the temperature 



