226 
PEOFESSOE FOECHHAMMEE ON THE COMPOSITION 
probably from animalculae and diatomacese. The place was not far from Victoria Land, 
at some distance from Coulman Island. It contained only 15-598 salt. The third, from 
65° 57' S. lat., 164° 37' E. long., had the surprising quantity of salt 37‘513 per 1000. 
The mean of these three observations is 27’225 per 1000 ; but this mean number is of 
very little consequence, being derived from numbers differing so greatly. It is, however, 
very surprising that water from the neighbourhood of the supposed Antarctic continent 
should have a salinity higher than any one found in the south equatorial regions of the 
Atlantic, and only be exceeded by a single one in the North Atlantic regions. I am 
sure that no material fault exists in the analysis, and this curious fact must thus remain 
unexplained until repeated observations in that region shall procure us further informa- 
tion. Should the observation be proved to be correct, it would render the existence of 
a “ Gulf-stream ” in the Antarctic zone very probable. There is still another peculiarity 
in these observations which deserves attention, viz. the great proportion of sulphuric 
acid to chlorine. In the water in the neighbourhood of Coulman’s Island it is 
12-47 : 100, and in that from 65° 57' S. lat. 12 - 55 : 100, while in the whole ocean it 
is as 11‘89 : 100. This might depend upon the very pronounced volcanic character of 
the Antarctic continent. There is still one question to be discussed with respect to 
the Antarctic Sea, how it is to enter into the mean numbers of the whole ocean. The 
observation from the neighbourhood of Coulman’s Island must be rejected, because it is 
too near the land, and we have no corresponding observations from the open Antarctic 
Ocean. Its high coefficient (1*861) shows the great influence of the neighbouring land. 
The observation from 65° 57' S. lat. must also be rejected as doubtful; there remains 
only the observation from the neighbourhood of the great ice-barrier, and I have taken 
that for the mean of the Antarctic region. 
General Results of the preceding investigation. 
If we except the North Sea, the Kattegat, Sound, and Baltic, the Mediterranean and 
Black Sea, the Caribbean Sea and the Bed Sea, which have all the characters of bays 
of the great ocean, the mean numbers are the following : — 
Sea-water. 
Chlorine. 
Sulphuric acid. 
Lime. 
Magnesia. 
All salts. 
Coefficient. 
1000 
18-999 
2-258 
0-556 
2-096 
34*404 
1*812 
100 
11-88 
2-93 
11-03 
Equivalents 
429 
45 
16 
82 
Thus it is evident that sea-water in its totality is as little a chemical compound as the 
atmospheric air ; that it is composed of solutions of different chemical compounds ; that 
it is neutral, because it everywhere in the atmosphere finds carbonic acid to neutralize 
its bases, and everywhere on its bottom and shores finds carbonate of lime to neutralize 
any prevailing strong acid ; that, lastly, the great stability of its composition depends 
upon its enormous mass and its constant motion, which occasions that any local varia- 
tion is evanescent compared to the whole quantity of salt. 
