H. R. Mill in 1885 devised a self-locking arrangement by which the bottle once closed was automatically locked and rendered watertight; H. L. Ekman made further improvements; and, finally, O. Pettersson and F. Nansen perfected the instrument, adapting it not only for enclosing a portion of water at any desired depth, but by a series of concentric divisions insulating in the central compartment water at the temperature it had at the moment of collection. By means of a weight dropped along the line the water-bottle can be shut and a sample enclosed at any desired depth. The use of a sliding weight is not recommended in depths much exceeding 200 fathoms on account of the time required and the risk of the line sagging at a low angle and so stopping the weight. In deep water the closing mechanism is usually actuated by a screw propeller which begins to work when the line is being hauled in and can be set so as to close the water-bottle in a very few fathoms. A small but heavy water-bottle has been devised by Martin Knudsen, provided with a pressure gauge or bathometer, by which samples may be collected from any moderate depth down to about 100 fathoms, on board a vessel going at full speed. This has made it possible to obtain many samples from moderate depths along a long line in a very short space of time. Sigsbee’s small water-bottle on the double valve principle actuated by a propeller requires extremely skilful handling to enable it to give good results.
As yet it is only possible to speak with confidence of the vertical distribution of salinity in the seas surrounding Europe, where there is a general increase of salinity with depth. For the open ocean the only quite trustworthy results are those obtained by the prince of Monaco in the North Atlantic, and by the recent Antarctic expeditions in the South Atlantic and South Indian Oceans. The observations made on the “Challenger” and “Gazelle,” though enabling some perfectly sound general conclusions to be drawn, require to be supplemented. It appears, as J. Y. Buchanan pointed out in 1876, that the great contrasts in surface salinity between the tropical maxima and the equatorial minima give place at the moderate depth of 200 fathoms to a practically uniform salinity in all parts of the ocean.
In the North Atlantic a strong submarine current flowing outward from the Mediterranean leaves the Strait of Gibraltar with a salinity of 38 per mille, and can be traced as far as Madeira and the Bay of Biscay in depths of from 600 to 2800 fathoms, still with a salinity of 35·6 per mille, whereas off the Azores at equal depths the salinity is from 0·5 to 0·7 per mille less. In the tropical and subtropical belts of the Atlantic and Indian Oceans south of the equator the salinity diminishes rapidly from the surface downwards, and at 500 fathoms reaches a minimum of 34·3 or 34·4 per mille; after that it increases again to 800 fathoms, where it is almost 34·7 or 34·8, and this salinity holds good to the bottom, even to the greatest depths, as was first shown by the “Gauss” and afterwards by the “Planet” between Durban and Ceylon.
Our knowledge of the Pacific in this respect is still very imperfect, but it appears to be less salt than the other oceans at depths below 800 fathoms, as on the surface, the salinity at considerable depths being 34·6 to 34·7 in the western part of the ocean, and about 34·4 to 34·5 in the eastern, so that, although the data are by no means satisfactory, it is impossible to assign a mass-salinity of more than 34·7 per mille for the whole body of Pacific water.
The causes of difference of salinity are mainly meteorological. The belt of equatorial minimum salinity corresponds with the excessively rainy belt of calms and of the equatorial counter-current, the salinity diminishing towards the east. The tropical maxima of salinity on the poleward side of the trade-winds coincide with the regions of minimum rainfall, high temperature, strong winds and consequently of maximum evaporation. Evaporation is naturally greatest in the enclosed seas of the nearly rainless subtropical zone such as the Mediterranean and Red Sea. Where the evaporation is at a minimum, the inflow of rivers from a large continental area and the precipitation from the atmosphere at a maximum, there is necessarily the greatest dilution of the sea-water, the Baltic and the Arctic Sea being conspicuous examples.
Temperature of the Oceans.—There is no difficulty in observing the temperature of the surface of the sea on board ship, the only precautions required being to draw the water in a bucket which has not been heated in the sun in summer or exposed to frost in winter, to draw it well forward of any discharge pipes of the steamer, to place it in the shade on deck, insert the thermometer immediately and make the reading without delay. The measurement of temperature in the depths, unless a high-speed water-bottle be used, involves stopping the ship and employing thermometers of special construction. Many forms have been tried, but only three types are in general use. The first is the slow-action thermometer which was originally used with good effect by de Saussure in the Mediterranean in 1780. He covered the bulb of the thermometer with layers of non-conducting material and left it immersed at the desired depth for a very long time to enable it to take the temperature of its surroundings. When brought up again the thermometer retained its temperature so long that there was ample time to take a correct reading. Since 1870 thermometers on this principle have been in use for regular observations at German coast and light-ship stations. Following the suggestion of Cavendish, Irving made observations of deep temperature on Phipps’s Spitsbergen voyage of 1773 with a valved water-bottle, insulated by non-conducting material. A similar instrument gave excellent results in the hands of E. von Lenz on Kotzebue’s second voyage of circumnavigation in 1823–1826. The last elaboration of the insulated slip water-bottle by Ekman, Nansen and Pettersson has produced an instrument of great perfection, in which the insulation is effected by layers of water between a series of concentric ebonite cylinders, all of which are closed both above and below when the apparatus encloses a sample, and each of which in turn must be warmed considerably before there is any rise of temperature in the chamber within. This can be used with certainty to ·02° C. for water down to 250 fathoms, after taking account of the slight disturbance produced by the expansion of the greatly compressed deep water.
The second form of deep-sea thermometer is the self-registering maximum and minimum on James Six’s principle. These instruments must be constructed with the greatest care, but when well made in accordance with J. Y. Buchanan’s large model they can be trusted to give a good account of the vertical distribution of temperature, provided the water grows cooler as the depth increases. They would act equally well if the water grew continually warmer as the depth increases, but they cannot give an exact account of a temperature inversion such as is produced when layers of warmer and colder water alternate.
The third form is the outflow or reversing thermometer, first introduced by Aimé, who used a very inconvenient form in the Mediterranean in 1841–1845, but greatly improved and simplified by Negretti and Zambra in 1875. The principle is to have a constriction in the tube above the bulb so proportioned that when the instrument is upright it acts in every way as an ordinary mercurial thermometer, but when it is inverted the thread of mercury breaks at the constriction, and the portion above the point runs down the now reversed tube and remains there as a measure of the temperature at the moment of turning over. For convenience in reading, the tube is graduated inverted, and when it is restored to its original position the mercury thread joins again and it acts as before. Various modifications of this form of thermometer have been made by Chabaud of Paris and others. It has the advantage over the thermometer on Six’s principle that, being filled with mercury, it does not require such long immersion to take the temperature of the water. A correction has, of course, to be made for the expansion or contraction of the mercury thread if the temperature of reading differs much from that of reversing. Magnaghi introduced a convenient method of inverting the thermometer by means of a propeller actuated on beginning to heave in the line, and this form is used for all work at great depths. For shallow water greater precision and certainty are obtained by using a lever
