table it is not to be supposed that a dilute solution like sea-water contains all the ingredients thus arbitrarily combined. There must be considerable dissociation of molecules, and as a first approximation it may be taken that of 10 molecules of most of the components about 9 (or in the case of magnesium sulphate 5) have been separated into their ions, and that it is only during slow concentration as in a natural saline that the ions combine to produce the various salts in the proportions set out in the above table. One can look on sea-water as a mixture of very dilute solutions of particular salts, each one of which after the lapse of sufficient time fills the whole space as if the other constituents did not exist, and this interdiffusion accounts easily for the uniformity of composition in the sea-water throughout the whole ocean, the only appreciable difference from point to point being the salinity or degree of concentration of the mixed solutions.
The origin of the salt of the sea is attributed by some modern authorities entirely to the washing out of salts from the land by rain and rivers and the gradual concentration by evaporation in the oceans, and some (e.g. J. Joly) go so far as to base a calculation of the age of the earth on the assumption that the ocean was originally filled with fresh water. This hypothesis, however, does not accord with the theory of the development of the earth from the state of a sphere of molten rock surrounded by an atmosphere of gaseous metals by which the first-formed clouds of aqueous vapour must have been absorbed. The great similarity between the salts of the ocean and the gaseous products of volcanic eruptions at the present time, rich in chlorides and sulphates of all kinds, is a strong argument for the ocean having been salt from the beginning. Two other facts are totally opposed to the origin of all the salinity of the oceans from the concentration of the washings of the land. The proportions of the salts of river and sea-water are quite different, as Julius Roth shows thus:—
| Carbonates. | Sulphates. | Chlorides. | |
| River water | 80 | 13 | 7 |
| Sea water | 0·2 | 10 | 89 |
The salts of salt lakes which have been formed in the areas of internal drainage in the hearts of the continents by the evaporation of river water are entirely different in composition from those of the sea, as the existence of the numerous natron and bitter lakes shows. Magnesium sulphate amounts to 4·7% of the total salts of sea-water according to Dittmar, but to 23·6% of the salts of the Caspian according to Lebedinzeff; in the ocean magnesium chloride amounts to 10·9% of the total salts, in the Caspian only to 4·5%; on the other hand calcium sulphate in the ocean amounts to 3·6%, in the Caspian to 6·9%. This disparity makes it extremely difficult to view ocean water as merely a watery extract of the salts existing in the rocks of the land.
The determination of salinity was formerly carried out by evaporating a weighed quantity of sea-water to dryness and weighing the residue. Forchhammer, however, pointed out that this method gave inexact and variable results, as in the act of evaporating to dryness hydrochloric acid is given off as the temperature is raised to expel the last of the water, and Tornöe found that carbonic acid was also liberated and that the loss of both acids was very variable. Tornöe vainly attempted to apply a correction for this loss by calculation; and subsequently S. P. L. Sörensen and Martin Knudsen after a careful investigation decided to abandon the old definition of salinity as the sum of all the dissolved solids in sea-water and to substitute for it the weight of the dissolved solids in 1000 parts by weight of sea-water on the assumption that all the bromine is replaced by its equivalent of chlorine, all the carbonate converted into oxide and the organic matter burnt. The advantage of the new definition lies in the fact that the estimation of the chlorine (or rather of the total halogen expressed as chlorine) is sufficient to determine the salinity by a very simple operation. According to Knudsen the salinity is given in weight per thousand parts by the expression S＝0·030 + 1·8050 Cl where S is the salinity and Cl the amount of total halogen in a sample. Such a simple formula is only possible because the salts of sea-water are of such uniform composition throughout the Whole ocean that the chlorine bears a constant ratio to the total salinity as newly defined whatever the degree of concentration. This definition was adopted by the International Council for the Study of the Sea in 1902, and it has since been very widely accepted.
Besides the determination of salinity by titration of the chlorides, the method of determination by the specific gravity of the sea-water is still often used. In the laboratory the specific gravity is determined in a pyknometer by actual weighing, and on board ship by the use of an areometer or hydrometer. Three types of areometer are in use: (1) the ordinary hydrometer of invariable weight with a direct reading scale, a set of from five to ten being necessary to cover the range of specific gravity from 1·000 to 1·031 so as to take account of sea-water of all possible salinities; (2) the “Challenger” type of areometer designed by J. Y. Buchanan, which has an arbitrary scale and can be varied in weight by placing small metal rings on the stem so as to depress the scale to any desired depth in sea-water of any salinity, the specific gravity being calculated for each reading by dividing the total weight by the immersed volume; (3) the total immersion areometer, which has no scale and the weight of which can be adjusted so that the instrument can be brought so exactly to the specific gravity of the water sample that it remains immersed, neither floating nor sinking; this has the advantage of eliminating the effects of surface tension and in Fridtjof Nansen’s pattern is capable of great precision.
In all areometer Work it is necessary to ascertain the temperature of the water sample under examination with great exactness, as the volume of the areometer as well as the specific gravity of the water varies with temperature. All determinations must accordingly be reduced to standard temperature for comparison. Following the practice of J. Y. Buchanan on the “Challenger” it has been usual for British investigators to calculate specific gravities for sea-water at 60° F. compared with pure water at the maximum density point (39·2°) as unity. On the continent of Europe it has been more usual to take both at 17·5° C. (63·5° F.), which is expressed as “S17·517·5”, but for pyknometer work in all countries where the sample is cooled to 32° F. before weighing and pure water at 39·2° taken as unity the expression is (0°/4°). On the authority of the first meeting of the International Conference for the Study of the Northern European Seas at Stockholm in 1899 Martin Knudsen, assisted by Karl Forch and S. P. L. Sörensen, carried out a careful investigation of the relation between the amount of chlorine, the total salinity and the specific gravity of sea-water of different strengths including an entirely new determination of the thermal expansion of sea-water. The results are published in his Hydrographical Tables in a convenient form for use.
The relations between the various conditions are set forth in the following equations where σ0 signifies the specific gravity of the sea-water in question at 0° C., the standard at 4° being taken as 1000, S the salinity and Cl the chlorine, both expressed in parts by weight per mille.
| (1) σ0 = −0·093 + 0·8149 S − 0·000482 S2 + 0·0000068 S3 |
| (2) σ0 = −0·069 + 1·4708 Cl − 0·00157 Cl2 + 0·0000398 Cl3 |
| (3) S = 0·030 + 1·8050 Cl. |
The temperature of maximum density of sea-water of any specific gravity was found by Knudsen to be given with sufficient accuracy for all practical purposes by the formula θ＝3·95 − 0·266σ0, where θ is the temperature of maximum density in degrees centigrade. The temperature of maximum density is lower as the concentration of the sea-water is greater, as is shown in the following table:—
| Salinity per mile | 0 | 10 | 20 | 30 | 35 | 40 |
| Temperature θ° C | 3·95° | 1·86° | − 0·31° | − 2·47° | − 3·52° | − 4·54° |
| Density σθ | 0·00° | 8·18° | 16·07° | 24·15° | 28·22° | 32·32° |
