of the Fishery Board for Scotland. 



283 



that the results of examination of 3000 samples give full confidence in 

 the correctness of these tables, based on Prof. Pettersson's data, for all 

 waters likely to be met with in the North Atlantic. The differences 

 of salinity given in column 10 of Table III., corresponding in 

 magnitude to those obtained by most other observers, have apparently no 

 chemical significance, and must be taken as expressing the accumulated 

 errors in collecting the samples, storing them, and making the two 

 determinations. They are, in any case, well within the point of attaining 

 any geographical meaning. 



Column 1 1 of Table III. gives the densities at 15° C. calculated by 

 working backwards from column 9 with Table XII. of the 'Jackal' 

 report, and from column 11 the specific gravities in situ have been com- 

 puted, using the observed temperatures (T.) and Table X. of the former 

 paper (col. 12). 



Notwithstanding the failure of all attempts hitherto made to find 

 chemical differences in the composition of sea-water, now apparently 

 made final by the identical values obtained for salinity from chlorines and 

 densities, I thought it possible * that waters derived largely from sea-water 

 ice might contain some measurable excess of sulphates compared with 

 waters coming from low latitudes, and that at least some indication might 

 thus be given of the place of origin of any given sample. A large 

 number of the surface samples from the Atlantic have been thus 

 examined, with results, mostly negative, which need not be discussed 

 here. Determinations of the sulphates in a number of the 1 Eesearch ' 

 samples were made along with others, and the values obtained are to be 

 found in column 13 of Table III. Fifty c.c. of the sample were taken in 

 each case, the sulphates precipitated with barium chloride in the usual 

 way, and left to stand overnight. Special care was taken to wash the 

 precipitates with boiling water until the combined filtrates of a whole 

 batch (usually numbering about eight) gave no perceptible turbidity with 

 silver nitrate. About half a dozen samples were done in duplicate, the 

 results agreeing closely in every case. The numbers in column 14, which 

 are those of column 13 multiplied by 100, and divided by those of column 

 6, show no variation amongst themselves which can be taken to repre- 

 sent actual differences of composition, but as most of them tend towards 

 the superior limit usually assigned to this fraction, it may be well to add 

 that they agree with the values obtained for the great bulk of the surface 

 samples. 



It appears from this that we must look to the analysis of dissolved 

 gases for any hope of being able to identify sea-waters by self-borne 

 labels, and one cannot but regret that the ' Research ' was not supplied 

 with tubes for collecting gas samples. 



Results. 



The general distribution of temperature and salinity disclosed by these 

 observations is easily described. At the surface the isothermal of 12° C. 

 passes in a W.N.W. to E.S.E. direction just north of the Ridge, forming 

 an almost straight line to the Orkney Islands. The general map for the 

 same period, not yet quite completed, shows that east of the Orkneys it 

 sweeps northward to the Norwegian coast, which it cuts in about 70° N. 

 lat., just north of the Lofotens. The line of 11° C, according to the 

 same chart, is nearly parallel to that of 12°, passing through the island of 

 Sudero in the Faeroes, and bending north to the east of the group, showing 

 a cold axis close inshore. Below the surface, where observations are 

 * See K. Rordam, Meddelelser om Gronland, xvii. p. 234. 



